!!"#$%&%'$!(()*+$,--./(0$#-1./2($2($3*./2(*+$#-1)4/.5$"*6$!"#$%#&'(&)(*&+,(*#&(-&./*#&*0/%&1(-23/)34-(5-#%%&1/*0(,*&$,*0(-6%&4#-7/%%/()&

8()*$.*9&",:/)$;/,<#',&&

BLIND ORACLES: REGULATING INTELLIGENCE PROCESSING, ANALYSIS, AND

VERIFICATION IN WARTIME AERIAL STRIKES

Asaf Lubin* =!"#$%&#&

The likelihood of catastrophic effects from faulty intelligence assessments is perhaps most visible in the context of wartime aerial attacks. Modern history is filled with examples of intelligence errors that resulted in calamitous air campaigns. In this paper I look at five such cases, spanning various historical periods, geographical zones, and belligerent parties: the Battle of Monte Cassino during WWII (1944), the U.S. raid on a psychiatric hospital during the invasion into Grenada (1983); the Allied Forces bombing of the Chinese embassy in Belgrade (1999); the German attack on tanker-trucks in the Kunduz Province of Afghanistan (2009); the Israeli Gaza beach attack as part of Operation Protective Edge (2014).

Examining these case studies, the paper makes the claim that faults in wartime intelligence production are not as inevitable as often presumed and that it is for a lack of specific regulation within the treatises of International

&* !"#$%&'($)*+,-$,&$'-$%&&./,'01$2".(1&&."$.($)'3$'0$4-5,'-'$6-,71"&,08$9'*"1"$:/;..<$

.($ )'3=$ '$ >'/*<08$ %&&./,'01$ '0$ 0;1$ ?1"@A'-$ B<1,-$ C1-01"$ (."$ 4-01"-10$ '-5$ :./,108$ '0$D'"7'"5$6-,71"&,08=$ '$E,&,0,-F$>1<<.3$ '0$ 0;1$ 4-(."A'0,.-$ :./,108$2".G1/0$ .($H'<1$)'3$:/;..<=$'$E,&,0,-F$:/;.<'"$'0$0;1$>151"A'--$C8+1"&1/*",08$C1-01"$'0$D1+"13$6-,71"&,08$.($ I1"*&'<1A=$ '$ >1<<.3$ '0$ 0;1$ C1-01"$ (."$ %JJ<,15$ C8+1"&1/*",08$ K1&1'"/;$ '0$ 4-5,'-'$6-,71"&,08=$'-5$'$E,&,0,-F$>1<<.3$'0$0;1$L1+"'&@'$M.71"-'-/1$'-5$N1/;-.<.F8$C1-01"$'0$0;1$6-,71"&,08$.($L1+"'&@'. This paper benefited from the thought provoking comments of W. Michael Reisman, Robert Chesney, Christophe Bondy, Marco Sassoli, Marko Milanovic, Claus Kreß, Leila Nadya Sadat, David Sloss, Craig Martin, Milena Sterio, and my students at the Yale College Seminar CSMY 370: Espionage and International Law. The paper was workshopped at the Second Graduate Public Law Conference at the University of Texas at Austin, at the Tenth International Graduate Legal Research Conference at Kings College London (where it was awarded the Kennedy Prize for best conference presentation), at the Fifth Annual Doctoral Scholarship Conference at Yale Law School, at the 12th European Society of International Law Annual Conference, at the 7th American Society of International Law Midyear Research Forum, at a Junior Exchange Workshop at Notre Dame Law School, at the 2020 ASIL Midwest Interest Group Virtual Works in Progress Conference, and at the AALS 2021 Annual Meeting Section on National Security Law Works-in-Progress Session. I wish to thank all participants in all those workshops for their terrific comments and thoughts which benefited this paper greatly. Finally, I wish to thank Harry Zheng for excellent research assistance.

>& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

Humanitarian Law (IHL) that they occur at the rate that they do. Tribunals and military manuals guide us to rely on the "reasonable commander" test in determining the lawfulness of a particular strike. Yet, in the process we overlook the fact that any reasonable commander will turn to her "reasonable intelligence analyst"—the contours of this standard are conspicuously under-defined. This paper takes a first step at proposing such an international standard, a new duty of care, based on both historical analysis and emerging soft-law norms and best practices. In so doing the paper proposes a path forward for addressing the accountability gap that permeates contemporary IHL as it relates to wartime errors and mistakes. B%!'()*+)8*,#(,#"& C)*-(',.*/()&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&D C< C)*#""/5#).#&!-(',.*/()&/)&CEF&G&=)&HI#-I/#1&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&J

=<&B0#&K/50*&*(&L4M&/)&N$-&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&AO B.&B0#&F$1&P(I#-)/)5&C)*#""/5#).#&H4#-$*/()%&/)&N$-*/7#&<<<<<<<<<<<<<<<<<<&AQ

CC< 8$%#&L*,'/#%9&R$,"*M&C)*#""/5#).#&/)&=#-/$"&L*-/2#%&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&>A A. Case Selection Criteria&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&>> B.&=)$"M%/%&(S&8$%#&L*,'/#%&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&>D

A< B0#&T$**"#&(S&U()*#&8$%%/)(&VR#:-,$-M&AWX&AJOOY&<<<<<<<<<<<<<<<<<<<&>D 2. Operation “Urgent Fury” and the Bombing of the Richmond Hill Mental Hospital (October 25, 1983)&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&>Q D< B0#&=**$.2&()&*0#&80/)#%#&Z7:$%%M&VU$M&[X&AJJJY&<<<<<<<<<<<<<<<<<&>J O< P#-7$)&=**$.2&()&B$)2#-&B-,.2%&(,*%/'#& *0#&\/""$5#&(S&E$]/&L$20/&^#':M&/)&_,)',`&!-(I/).#&(S&=S50$)/%*$)&VL#4*#7:#-&OX&>aaJY&<<<<&D> W< H4#-$*/()& b!-(*#.*/I#& Z'5#c& $)'& *0#&P$`$& T#$.0&=**$.2& V@,"M& AQX&>aAOY&&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&DW

CCC< K#.()S/5,-/)5&C)*#""/5#).#&/)&N$-&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&Dd =<&U$2/)5&*0#&8$%#&S(-&$&e#1&^,*M&(S&8$-#&S(-&C)*#""/5#).#&<<<<<<<<<<<<<<<<&Oa

A< =''-#%%/)5&*0#&=..(,)*$:/"/*M&\$.,,7&S(-&N$-*/7#&C)*#""/5#).#&<&Oa >< Z)0$)./)5&C)*#-)$*/()$"&F$16%&Zf4-#%%/I#&\$",#&<<<<<<<<<<<<<<<<<<<<&OA D< ̂ #I#"(4/)5&U#.0$)/%7&S(-&K#'-#%%&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&OD

T<&bB0#&K#$%()$:"#&C)*#""/5#).#&=)$"M%*c&L*$)'$-'&<<<<<<<<<<<<<<<<<<<<<<<<<<<<&OO A< =&8()*#f*3T$%#'&^,*M&(S&8$-#&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&OW >< L*-,.*,-$"&$)'&!-(.#',-$"&^,*/#%&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&OQ

8<&R/**/)5&*0#&e#1&^,*M&(S&8$-#&N/*0/)&Zf/%*/)5&CEF&<<<<<<<<<<<<<<<<<<<<<<<<&W> 8().",%/()&<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<&WO

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& D&

C,#$*-.&#/*,&

In November 2019 yet another incident occurred in Gaza. It seemed like

another round of violence between the Israeli military and the Palestinian Islamic Jihad (PIJ). However, what began as a routine approval of a targeting sequence against a PIJ military training complex in the Palestinian town of Dir al-Balah quickly turned into a disaster. A series of intelligence errors led to a mistake in identification that resulted in the death of a family of eight, including two toddlers.1 Here are but some of the blunders: (1) the site in question was originally designated as an “infrastructure target” but in the “period since it was approved as a target, the changes at the complex were not looked into to determine if it still served as an Islamic Jihad site;2 (2) as was later confirmed, “at no stage was the area checked for the presence of civilians”;3 (3) information as to the fact that a family of herders used the series of shacks at the site as a home for quite some time prior to the incident was never identified or examined;4 (4) Immediately following the incident the Israeli Arabic-language spokesperson falsely claimed that the operation targeted Abu Malhous who was said to be in charge of the PIJ’s rocket squadrons in central Gaza. In reality, no such individual was ever known to the intelligence community and it seems that the false statement was based on “unreliable information shared on social media.”5 Israeli officials later claimed that this was “an innocent mistake, while admitting the way the incident was handled and made public was unprofessional.”6

It is common for militaries to brush such incidents aside as an unintended mistake amidst the tormenting “fog of war.” But it is a short road from designating an incident as “unintentional” to the follow-up designation of “unavoidable.” From there, the path is made clear for the incident to be branded as “irreprehensible,” “comprehensible,” and ultimately “defensible.” Here lies the danger. To what extent is our willingness to tolerate certain

&!$!""#H'-,7$B*+.7,/;$'-5$I'/@$B;.*"8=$$%&'"()#*&+,#*-+).%#./#0)(()12#3)24.#5'6'#

7'+)(,# 8"+9"&%:# ;"# <4/=24.# .4"# >/=%"# ;'%# 3+?.,=$ D!!"#$%$ OL.7#$ PQ=$ RSPTU=$;00J&VWW333#;''"10X#/.AWA,55<1Y1'&0Y-13&WJ'<1&0,-,'-&W#J"1A,*AY,&"'1<,Y'"A8Y'5A,0&Y&0",@1Y0;'0Y@,<<15YJ'<1&0,-,'-Y('A,<8Y,-01-515Y(."Y1AJ08Y;.*&1YP#ZPRT[\Q]<0&^P_ST_PQ[`RQPZ#$

"$!""#H'-,7$B*+.7,/;$'-5$I'/@$B;.*"8=$@=.-'."-#$1."(()2"1A"B#!/A)'(#8"-)'#C=+/&%:#D"4)1-# $%&'"(E%# 0)(()12# /F# 5'6'# 7'+)(,=$ D!!"#$%& OL.7#$ PQ=$ RSPTU=&;00J&VWW333#;''"10X#/.AW,&"'1<Y-13&W#J"1A,*AY.*05'015Y,-01<<,F1-/1Y&./,'<YA15,'Y"*A."&Y+1;,-5Y,&"'1<Y&Y@,<<,-FY.(YF'X'Y('A,<8YP#ZP\PPSP#&$

#$$-#$$$$-#$%$$-#$&$$-#$

O& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

wartime errors, specifically those resulting from faults in the intelligence process, a serious failure of our contemporary laws of war? To what extent does our willingness to name them as mere “mistakes” result in a reaffirmation cycle that hinders moral and legal line drawing?

James Dawes once wrote that the very act of naming “produces an Other, establish hierarchies, enable surveillance, and institute violent binaries.”7 If naming produces violence, as Dawes suggests, then intelligence agencies are in the most violent of all naming industries. After all, the work of the wartime intelligence analyst is to continuously name, categorize, and label. In the morning, the analyst inserts geographical coordinates into a database indicating a military target based on certain signals intelligence (SIGINT).8 Right before lunch, the analyst produces a detailed memo on the side-dealings of some civilian farmer branding her with a new title: a “direct participant in hostilities.”9 By evening, the analyst employs predictive algorithms to identify a ring of youth susceptible to the appeal of terrorism marking them for future questioning and potential jail time.10

&'$I'A1&$K#$!'31&=$G'12=''2"B#H)/("1A"B#'1-#>=+'1#C)24.%#G'I=$PP$H!'#&I#)#&a&

D()!*&RPQ=$RPQ$OPTTTU#$($!""#"J2J#@FF#<'&2".:#<4"#K/1-=A.#/F#;'&#'1-#K)L)()'1#K'=%'().)"%#)1#$&'M=$D()!*&

K+,-$.&b!$/-$O!1/#$PP=$RSS\U=$;00J&VWW333#;"3#."FW"1J."0WRSS\WPRWPPW0'"F10W/.-5*/0Y3'"Y'-5Y/,7,<,'-Y/'&*'<0,1&Y,"'c$$Od9'-8$.($0;1$/,7,<,'-$/'&*'<0,1&$(".A$0;1$',"$3'"$.//*""15$5*",-F$ $6#:#$ '00'/@&$ 0'"F10,-F$ &1-,."$ 4"'c,$ <1'51"&#$N;1$6-,015$:0'01&$*&15$'-$*-&.*-5$0'"F10,-F$ A10;.5.<.F8$ 0;'0$ "1<,15$ .-$ ,-01"/1J0&$ .($ &'01<<,01$ J;.-1&$ '-5$ ,-'51c*'01$/."".+."'0,-F$ ,-01<<,F1-/1#$ N;*"'8'$ &'01<<,01$ J;.-1&$ J".7,51$ F1.Y/.."5,-'01&$ 0;'0$ '"1$'//*"'01$.-<8$0.$3,0;,-$'$.-1Y;*-5"15YA101"$O\RZY(..0U$"'5,*&e$0;1"1(."1=$0;1$6-,015$:0'01&$/.*<5$-.0$5101"A,-1$ 0;1$.",F,-$.($'$/'<<$ 0.$'$51F"11$.($'//*"'/8$F"1'01"$ 0;'-$'$\P=[SSY&c*'"1YA101"$'"1'#$$N;,&$(<'315$0'"F10,-F$&0"'01F8$3'&$/.AJ.*-515$+8$'$<'/@$.($1((1/0,71$'&&1&&A1-0$+.0;$J",."$0.$0;1$'00'/@&$.($0;1$J.01-0,'<$",&@&$0.$/,7,<,'-&$'-5$'(01"$0;1$'00'/@&$.($0;1,"$&*//1&&$'-5$*0,<,08#$$%<<$.($0;1$(,(08$'/@-.3<15F15$'00'/@&$0'"F10,-F$4"'c,$<1'51"&;,J$(',<15#$b;,<1$0;18$5,5$-.0$@,<<$'$&,-F<1$0'"F1015$,-5,7,5*'<=$0;1$&0",@1&$@,<<15$'-5$,-G*"15$5.X1-&$.($/,7,<,'-&#fU#$

)$!""#"J2J#2".0./.<$44$'55,0,.-'<$0.$0;1$M1-17'$C.-71-0,.-&$.($PT[T=$'-5$"1<'0,-F$0.$0;1$ 2".01/0,.-&$ .($ E,/0,A&$ .($ L.-Y4-01"-'0,.-'<$ %"A15$ C.-(<,/0&=$ PPRQ$ 6#L#N#:$ _ST=$%"0,/<1$P\$OPT``U$g;1"1,-'(01"V$%244h$OdN;1$/,7,<,'-$J.J*<'0,.-$'-5$,-5,7,5*'<$/,7,<,'-&$&;'<<$1-G.8$ F1-1"'<$ J".01/0,.-&$ 'F',-&0$ 0;1$ 5'-F1"&$ '",&,-F$ (".A$ A,<,0'"8$ .J1"'0,.-&###$ N;1$/,7,<,'-$J.J*<'0,.-$'&$&*/;=$'&$31<<$'&$,-5,7,5*'<$/,7,<,'-&=$&;'<<$-.0$+1$0;1$.+G1/0$.($'00'/@###$C,7,<,'-&$&;'<<$1-G.8$0;1$J".01/0,.-$'((."515$+8$0;,&$2'"0=$*-<1&&$'-5$(."$&*/;$0,A1$'&$0;18$0'@1$ '$ 5,"1/0$ J'"0$ ,-$ ;.&0,<,0,1&#fU#$ >."$ (*"0;1"$ '-'<8&,&$ .($ J'"0,/*<'"$ <1F'<$ 'A+,F*,0,1&$"1&*<0,-F$ (".A$ 0;,&$ F1-1"'<$ "*<1$ &11$ L+'.& 9#'%#"=& 4*$#"0"#$+1#& M(+2!*/#& 3*& $-#&

L3$+3*& 34& !+"#/$& 2!"$+/+0!$+3*& +*& D3.$+'+$+#.& (*2#"& 4*$#"*!$+3*!'&

D()!*+$!"+!*&)!5=&4CKC$ORSSTU$g;1"1,-'(01"V$4CKC$4-01"J"10,71$M*,5'-/1$.-$!2Dh#$!*$i""$D,"&/;'*F1$'-5$D'F'"$:;1X'(=$>/I#$%&'"(#N')(%#O'("%.)1)'1%#D"A'=%"#<4",#7).#

.4"# E<"&&/&)%.# O&/F)("E=$ D!!"#$%$ O9'8$ \P=$ RSP`U=$ ;00J&VWW333#;''"10X#/.AW,&"'1<Y-13&W#J"1A,*A#9%M%j4LkY,&"'1<YG',<&YJ'<1&0,-,'-&Y3;.Y(,0Y01"".",&0YJ".(,<1YP#Q[``[\`$Od4&"'1<$ ;'&$ '""1&015$ ;*-5"15&$ .($ 2'<1&0,-,'-&$ &,-/1$ 0;1$ <.-1Y3.<($ ,-0,('5'$ +1F'-$ ,-$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& W&

But our review wouldn’t be complete if we only focused on the way the intelligence community names its targets. We should also ask ourselves what language intelligence professionals use to label their own activities, successes and failures. What signals are sent by naming something a mere “mistake”? It is the language of risk analysis in war that interests me in this project and so we might find it useful to start by asking what is deemed an intelligence “mistake” in the first place? It is of course true that some faults are rooted in capacity limitations. Intelligence agency’s assessments are only as good as the budgetary resources at its disposal, the human resources that it employs, and the intelligence sources it had already cultivated. These limitations are outside the scope of this paper.

There is, however, another set of reasons for suboptimal intelligence. Intelligence production, assessment, and verification are all parts of a professional tradecraft. Like any trade, these activities can be executed to varying degrees of success by individuals with varying degrees of skill and expertise. These professionals, therefore, may at times execute their tasks with willful, wanton, or gross negligence. Intelligence agencies may or may not adopt internal policies to prevent such behavior and to investigate and punish it once it occurs. Nonetheless, from group think and other cognitive biases11 to more fundamental and systemic intelligence failures, there may be many preventable faults in intelligence production and dissemination that an intelligence agency may simply be unwilling or unable to control.12

&:1J01A+1"$RSPQ=$J'"0<8$+8$'-'<8X,-F$&./,'<$A15,'$J.&0&l$:,-/1$0;1$3'71$.($7,.<1-/1$O'<&.$@-.3-$'&$0;1$<.-1Y3.<($,-0,('5'U$+1F'-$,-$:1J01A+1"$RSPQ=$0;1$4&"'1<,$'*0;.",0,1&$5171<.J15$'-$ 1'"<8$ '<1"0$ &8&01A$ 0;'0$ '&&1&&1&$ 0;1$ <,@1<,;..5$ .($ 0;1$ ,-7.<71A1-0$ .($ ,-5,7,5*'<$2'<1&0,-,'-&$ ,-$ 01""."$ '00'/@&l$:..-$'(01"$ 0;1$+1F,--,-F$.($ 0;1$3'71$.($ '00'/@&=$ 4&"'1<,$'*0;.",0,1&$ '&&,F-15$ 5.X1-&$ .($ ,-01<<,F1-/1$ .((,/1"&$ 0.$ '&&1A+<1$ F1-1"'<,X15$ J".(,<1&$ .($dJ.01-0,'<$'00'/@1"&=f$'//."5,-F$0.$0;1$.((,/1"#$4-,0,'<<8=$0;"11$."$(.*"$F1-1"'<,X15$J".(,<1&$.($0;1$1'"<8$'00'/@1"&$31"1$/.-&0"*/015=$/.A+,-,-F$5'0'$J.,-0&$,-/<*5,-F$'F1=$<./'0,.-$.($"1&,51-/1$'-5$'-$'&&1&&A1-0$.($0;1$J&8/;.<.F,/'<$+*,<5Y*J$'-5$,-01-0,.-&$.($0;1$'00'/@1"&=$+'&15$.-$ 0;1$ ,-(."A'0,.-$'7',<'+<1$ 0.$ 0;1$'*0;.",0,1&#fU#$>."$(*"0;1"$"1'5,-F$&11$%&;<18$!11@&=$O&"-)A.)L"#31"+)"%=$PS[$E!#&)#&K#1#$PQRT$ORSPZU#$

!!$i-$0;1$J&8/;.<.F8$.($,-01<<,F1-/1$'-'<8&,&=$'-5$,-$J'"0,/*<'"$0;1$/.AA.-$/.F-,0,71$+,'&1&$ &;'"15$ +8$ ,-01<<,F1-/1$ J1"&.--1<=$ &11$ K+/-!"2& I#& D#(#"=& 2.6/-3'3,6& 34&

4*$#''+,#*/#&%*!'6.+.$OPTTTU#$$!"$!""B#"J2J=$The COMM’N ON THE INTELLIGENCE CAPABILITIES OF THE U.S. REGARDING

WEAPONS OF MASS DESTRUCTION 2 (Mar. 31, 2005), govinfo.library.unt.edu/wmd/report/wmd_report.pdf (“We conclude that the Intelligence Community was dead wrong in almost all of its pre-war judgments about Iraq's weapons of mass destruction. This was a major intelligence failure. Its principal causes were the Intelligence Community's inability to collect good information about Iraq's WMD programs, serious errors in analyzing what information it could gather, and a failure to make clear just how much of its analysis was based on assumptions, rather than good evidence. On a matter of this importance, we simply cannot afford failures of this magnitude. After a thorough review, the Commission found no indication that the Intelligence Community distorted the

Q& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

This could be dangerous. Once produced, the information may be disseminated to all parts of the military, to an air force pilot, an armor crewman, a ground commander, or a chief of staff. Those military practitioners will rely on that intelligence to calculate their strategic and operational decision-making in battle. This paper rests on the premise that if International Humanitarian Law (IHL) is truly interested in limiting the effects of armed conflict and mitigating human suffering, then it must provide better regulation of the military’s intelligence function, for it is those preliminary processes that may determine the outcomes of particular attacks and of the war as a whole.

The potential for cataclysmic consequences from faulty intelligence assessments is perhaps most stark in the context of wartime aerial strikes. Modern history is filled with examples of wartime intelligence assessment errors that led to calamitous air campaigns: from the battle of Monte Cassino in World War II through the bombing of the Chinese embassy in Belgrade in 1999, all the way through to the U.S. attack on a Doctors Without Borders hospital in Kunduz in 2015. In all of these examples, and in others presented

&evidence regarding Iraq's weapons of mass destruction. What the intelligence professionals told you about Saddam Hussein's programs was what they believed. They were simply wrong.”); McCann v. United Kingdom, App. No. 18984/91, ECtHR Grand Chamber, Judgment, paras. 188-189, 213 (1995) ("The Government further submitted that in examining the planning of the anti-terrorist operation it should be borne in mind that intelligence assessments are necessarily based on incomplete information since only fragments of the true picture will be known... In sum, having regard to the decision not to prevent the suspects from travelling into Gibraltar, to the failure of the authorities to make sufficient allowances for the possibility that their intelligence assessments might, in some respects at least, be erroneous and to the automatic recourse to lethal force when the soldiers opened fire, the Court is not persuaded that the killing of the three terrorists constituted the use of force which was no more than absolutely necessary in defence of persons from unlawful violence..."); THE PUBLIC COMM’N TO EXAMINE THE MARITIME INCIDENT OF MAY 31 2010, PART ONE, paras. 243-248 (2011) (comparing between the intelligence assessments conducted by British authorities in McCann and those conducted by Israeli Authorities in the Mavi Marmara case, determining that: “where British authorities were faulted for making assumptions that led to a sense of increased risk (i.e., not considering that their intelligence assessments that a car bombing was imminent might be wrong)... [i]n the present case, the risk was underappreciated and the limitations in the ROE with respect to the use of less-lethal weapons (range, areas of the body to be targeted, etc.), while put in place to limit injury to civilians, proved very restrictive considering the situation faced by the soldiers that fast-roped to the Mavi Marmara... The anticipation of and planning for “worst case” scenarios could have better prepared the soldiers for the situation to which they were exposed. In preparing exclusively for less violent scenarios, the danger from a legal perspective is that the soldiers might overreact when confronted with such unanticipated threats. However, and this should be emphasized, looking at the operation as a whole, that appears not to have happened, as the soldiers acted continually to distinguish the types of threat posed in different situations, and they even switched back and forth between lethal and less-lethal weapons to address those threats.”).$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& [&

throughout the paper, operations were launched on the basis of insufficient intelligence and on errors in processing, analysis, and verification.

One might have assumed that these repeated incidents would spark a lively debate on the regulatory frameworks surrounding intelligence production in times of war. Yet, the opposite is true: the laws of war remain moot on this question, with national security lawyers and human rights activists often apprehensive about delving into the murky business of regulating espionage.13 This is quite peculiar given that it is in the interest of the military commander, the belligerent parties, and the international community as a whole, that prescriptive processes be developed so that such risks could be mitigated by deterring suboptimal intelligence operations and conclusions. Nonetheless, given that intelligence is still considered a “dirty word” in international law,14 no one has yet set out to address the practice. As a result, military commanders are provided only a general directive, to take “all precautions feasible” to minimize, if not avoid, “incidental loss of civilian life, injury to civilians, or damage to civilian objects.”15 But this

&!#$4$'A$"1A,-515$+8$0;1$3."5&$.($2".(1&&."$%(&;11-$I.;-$K'5&'-$3;.=$3;,<1$0"8,-F$

0.$1mJ<',-$3;8$<1F'<$&/;.<'"&;,J$.-$0;1$,-01"-'0,.-'<$<'3$.($d,-01<<,F1-/1$'/0,7,0,1&f$;'&$+11-$&.$&/'"/1$;'5$&*FF1&015$0;'0V$dk&J,.-'F1$,&$-1,0;1"$<1F'<$-."$,<<1F'<$*-51"$,-01"-'0,.-'<$<'3#$k&J,.-'F1$1m,&0&$+10311-$0;1$01/0.-,/$J<'01&$.($<1F'<$&8&01A&###$31$&;.*<5$0.<1"'01$0;1$'A+,F*,0,1&$'-5$J'"'5.m1&$,-;1"1-0$,-$0;1$3."<5n&$&1/.-5$.<51&0$J".(1&&,.-#$%//1J0,-F$0;'0$1&J,.-'F1$,&$+18.-5$0;1$<'3$31$&;.*<5$A.71$.-$0.$.0;1"$J".G1/0&$o$3,0;$F"'/1#f$O&11$John Radsan, The Unresolved Equation of Espionage and International Law, 28 MICH. J. INT'L L. 595, 596-597 (2007)).$

!$$ See Simon Chesterman, Intelligence Cooperation in International Operations, in INTERNATIONAL INTELLIGENCE COOPERATION AND ACCOUNTABILITY 124, 124 (Hans Born et. al. eds., 2011) (“[e]ver since the United Nations deployed peacekeepers into conflict zones it has been necessary to have a deep understanding of the theatre of operations and parties to a conflict, yet intelligence was long regarded as a “dirty word” as the 1984 Peacekeeper’s Handbook put it; “military information” was the preferred euphemism... The prospect of the United Nations or any other international organisation developing an independent intelligence collection capacity is remote. This is due to the understandable wariness of the part of states about authorising a body to spy on them, though the United Nations itself has been reluctant to assume functions that might undermine its actual or perceived impartiality. At the same time, however, this position reflects a larger anomaly in the status of intelligence under international law as an activity commonly denounced but almost universally practised: empowering an international organisation to engage in espionage might give the lie to this example of diplomatic doublethink.”). See also Iñaki Navarrete & Russel Buchan, Out of the Legal Wilderness: Peacetime Espionage, International Law and the Existence of Customary Exceptions, 51 CORNELL INT’L L. J. 897, 932 (2019) (Navarrete & Buchan cite to a statement made by the Pakistani Agent speaking before the ICJ in the Jadhav case, that as far as the travaux préparatoires are concerned the drafters of the Vienna Convention on Consular Relations made no reference to espionage and thereby “accepted the fiction that there were no spies”. Navarrete & Buchan rely on this “bold remark” to reinforce the idea that “States have long regarded espionage as a ‘dirty word’.”).$

!%$ 2".0./.<$ 4$ '55,0,.-'<$ 0.$ 0;1$ M1-17'$ C.-71-0,.-&$ .($ PT[T=$ '-5$ "1<'0,-F$ 0.$ 0;1$

d& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

requirement, set out in Article 57(1) of the First Additional Protocol to the Geneva Conventions offers few, if any, concrete obligations on States in designing and implementing their intelligence apparatus.

This paper claims that faults in the production of intelligence for aerial operations are not always inevitable, and that it is for this lack of specific regulation within the treatises of IHL that such faults occur at the rate that they do. The principal pillars of IHL—distinction, necessity, proportionality, humanity, and precautions in attack—heavily depend on a forgotten supporting beam, that of an adequate, sufficient, and reliable stream of intelligence information.

The paper takes a first important step at defining the contours of a new duty of care, the “reasonable intelligence analyst” test. To develop this test, the paper considers both historical case studies as well as emerging soft-law norms and best practices. The paper not only looks to establish the nature and scope of this new duty but further explores its codification within existing IHL, which may be said to trigger state responsibility for failure to comply with it.

The paper proceeds in three parts. Part I begins with a brief overview of the law governing intelligence production in IHL and its existing limitations. It will particularly focus on two primary questions: (1) where do we derive the right to spy in times of war? and (2) what are the laws that govern the conduct of such intelligence operations?

Part II will then provide a historical analysis of five case studies of aerial campaigns where faults were identified to have occurred either at the collection and processing phases or at the analysis and dissemination phases of the intelligence cycle and where those faults resulted in significant loss of civilian life.

Part III will then synthesize the insights gleaned from this historical review and make the general case for a new framework that would govern liability for faulty intelligence processing, analysis, and verification during war. This part will be composed of three subsections. The first will make the case for a new standard by highlighting its potential role in addressing the accountability gap in wartime intelligence regulation, enhancing the expressive function of IHL, and developing mechanisms for redress to victims. The section subsection will introduce the new duty of care, the standard of “the reasonable intelligence analyst,” and define its underlying structure. The final subsection will propose how an informal rulemaking agenda could be adopted to further develop and expand the application of the new standard under existing IHL.

&

2".01/0,.-&$.($E,/0,A&$.($4-01"-'0,.-'<$%"A15$C.-(<,/0&=$PPRQ$6#L#N#:$\=$%"0,/<1$Q`OPU$OPT``U$g;1"1,-'(01"V$%24h#$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& J&

C<& C,#(''/0(,&()!$*-.&#/*,)/,)CEF)G)=,)H1($1/(2& This Part provides a brief primer on the legal regime that governs

intelligence collection in times of war. Others have provided more comprehensive histories of the development of this regime, and I do not want to repeat too much of that here.16 My account will instead spend more time on the limitations of the existing legal structure, which is the focus of this Article.

As a preliminary matter, I should perhaps clarify my choice of terminology. It is worth noting that: “no internationally recognized and workable definition of ‘intelligence collection’ exists,”17 as a matter of law, theory, or practice.18 In previous work I proposed a working definition of “intelligence operations” which is based, in part, on the idea that these operations involve the “passive gathering, analysis, verification, and dissemination, of information.”19 As I explained there, such a definition “encompasses all primary fields of intelligence gathering (or INTs as they are known in traditional spy parlance),”20 namely human, signals, visual, and open source intelligence. Intelligence operations occur in both war and peace. What will designate a particular piece of information as one that bears on a nation’s wartime preparedness and action, will depend on the consumer and their utilization of the information for wartime related policymaking.

Intelligence for wartime aerial targeting operations is a specific sub-sect of wartime intelligence. The targets of such attacks range in nature from a military base to a plant manufacturing arms, and from the home of a known member of an armed group to a broadcasting station spewing political propaganda and coded military instructions. Wartime aerial targeting intelligence traditionally covers six unique aspects and features:

1. GPS coordinates of the target and its physical location. For example, in the context of U.S. aerial operations against the Taliban “ground units used GPS target-locating and laser-target designating equipment to guide strike

&!&$ !""B# "J2J=$ I3-*& B+.-=& 4*$#"*!$+3*!'& )!5& !*2& k.0+3*!,#$ PR\YPQ\$ OPTTQU$

O/;'J01"$ [$ .($ 0;1$ +..@=$ 0,0<15V$ db'"$ '-5$ k&J,.-'F1f$ 51&/",+1&$ 0;1$ 17.<*0,.-$ .($ 0;1$,-01"-'0,.-'<$<'3$"1F,A1$&*"".*-5,-F$,-01<<,F1-/1$/.<<1/0,.-$'-5$&J8,-F$,-$0;1$<'3&$.($3'"#$N;1$C;'J01"$(./*&1&$.-$1&J,.-'F1$,-$3'"$.-$<'-5=$'0$&1'=$'-5$,-$0;1$',"=$'&$31<<$'&$0;1$"1F*<'0,.-&$ &*"".*-5,-F$ 3'"$ &J,1&$ '-5$ 0;1$ ,-01"".F'0,.-$ .($ /,7,<,'-&$ '-5$ 2ib&$ (."$,-01<<,F1-/1$J*"J.&1&U#$

!'$Sulmasy and Yoo, Counterintuitive: Intelligence Operations and International Law, 28 MICH. J. INT'L L. 625, 637 (2006). $

!($Social Theorist Michael Warner providew a relevant survey of different definitions that have been proposed over the years, to no avail, see: Michael Warner, Wanted: A Definition of "Intelligence", 6(3) STUDIES IN INTELLIGENCE 15 (2002).$

!)$See Asaf Lubin, Liberty to Spy, 61(1) HARV. INT’L. L. J. 185, 192 (2020). $"*$Id., at 193.$

Aa& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

aircraft and [precision-guided munitions] to targets.”21 2. Aerial footage of the target and its surrounding. Often targeting

operations rely on real-time surveillance (using unmanned aerial vehicles, or UAVs, on-ground informants, pre-installed surveillance devices, or a combination thereof) to accompany the archived materials. Together the images produced serve to establish certain “patterns of behavior/life” surrounding the target.22

3. The primary functions of the target. When the target is an individual intelligence will be collected to verify the “membership” of that individual in an organized armed group or his or her “direct participation in hostilities,” as these terms are defined under IHL.23 For structures and other physical objects, intelligence is used to verify the targets’ specific military objective (distinguishing between those which are military objectives by their nature and location and those which are military objectives by their purpose or use).24 This information helps confirm, with reasonable certainty, that the targeting of these objects will not violate the principle of distinction.25

4. The existence of “sensitive sites” surrounding the target. These would be objects that receive special protections under IHL such as: hospitals and other medical facilities and vehicles, educational facilities, cultural and religious sites and property, dense civilian locales and dwellings, and other objects that serve the civilian population’s humanitarian needs (water

&"!$ >"'-&$ i&,-F'$ a$ 9'"@$ K.."5'=$ From Douhet to Drones, Air Warfare, and the

Evolution of Targeting, in TARGETING: THE CHALLENGES OF MODERN WARFARE 27, 60 (Ducheine, Schmitt, & Osinga eds., 2016).$

""$!""#"J2J#)'*"1-0$M,&1<=$<4"#O&)1A)?("#/F#O&/?/&.)/1'().,#)1#.4"#C=("%#5/L"&1)12#.4"#K/1-=A.#/F#>/%.)().)"%#=1-"&#$1."&1'.)/1'(#>=+'1).'&)'1#G'I:#$1."&1'.)/1'(#3P?"&.#8"".)12#QQRQS# N=1"# QTUV=$ 4CKC=$ ``$ O:1J0#$ \=$ RSPZU=$;00J&VWW333#,/"/#."FW1-W5./*A1-0W,-01"-'0,.-'<Y1mJ1"0YA110,-FY"1J."0YJ",-/,J<1YJ".J."0,.-'<,08$g;1"1,-'(01"V$4CKC$kmJ1"0$9110,-F$K1J."0h$O'&$J'"0$.($'$03.Y5'8$".*-50'+<1$171-0$/.Y."F'-,X15$+8$0;1$4CKC$'-5$$,-7.<7,-F$RS$,-01"-'0,.-'<$1mJ1"0&$.-$0;1$/.-5*/0$.($;.&0,<,0,1&=$.-1$1mJ1"0$A'51$"1(1"1-/1$0.$d'55,0,.-'<$'&&1&&A1-0&f$,-$0;1$(."A$.($dJ'001"-&$.($<,(1$'-'<8&,&f$0;'0$'"1$0"'5,0,.-'<<8$"1/.AA1-515$,-$."51"$0.$5101"A,-1$0;1$",&@$(".A$'-8$J.01-0,'<$0'"F10$';1'5$.($'*0;.",X,-F$'1",'<$0'"F10,-F$.J1"'0,.-&U#$$$

"#$!""#%=?&'#-.01$T$'-5$'//.AJ'-8,-F$01m0#$"$$%24=$%=?&'#-.01$PQ=$%"0#$QR$OdC,7,<,'-$.+G1/0&$&;'<<$-.0$+1$0;1$.+G1/0$.($'00'/@$."$

"1J",&'<&#$C,7,<,'-$.+G1/0&$'"1$'<<$.+G1/0&$3;,/;$'"1$-.0$A,<,0'"8$.+G1/0,71&###$4-$&.$('"$'&$.+G1/0&$'"1$/.-/1"-15$A,<,0'"8$.+G1/0,71&$'"1$<,A,015$0.$0;.&1$.+G1/0&$3;,/;$+8$0;1,"$-'0*"1=$<./'0,.-=$J*"J.&1$."$*&1$A'@1$'-$1((1/0,71$/.-0",+*0,.-$0.$A,<,0'"8$'/0,.-$'-5$3;.&1$0.0'<$."$J'"0,'<$51&0"*/0,.-=$/'J0*"1=$."$-1*0"'<,X'0,.-=$ ,-$ 0;1$/,"/*A&0'-/1$"*<,-F$'0$ 0;1$ 0,A1=$.((1"&$'$51(,-,01$A,<,0'"8$'57'-0'F1#fU#$

"%$ $-J# '0$%"0,/<1$ [Z$ Od4-$ ."51"$ 0.$ 1-&*"1$ "1&J1/0$ (."$ '-5$ J".01/0,.-$ .($ 0;1$ /,7,<,'-$J.J*<'0,.-$ '-5$ /,7,<,'-$ .+G1/0&=$ 0;1$ J'"0,1&$ 0.$ 0;1$ /.-(<,/0$ &;'<<$ '0$ '<<$ 0,A1&$ 5,&0,-F*,&;$+10311-$0;1$/,7,<,'-$J.J*<'0,.-$'-5$/.A+'0'-0&$'-5$+10311-$/,7,<,'-$.+G1/0&$'-5$A,<,0'"8$.+G1/0,71&$'-5$'//."5,-F<8$&;'<<$5,"1/0$0;1,"$.J1"'0,.-&$.-<8$'F',-&0$A,<,0'"8$.+G1/0,71&#fU#$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& AA&

reservoirs, power stations, large food factories, U.N. facilities, and refugee camps).26

5. The military advantage to be gained from attacking the target. Lawful attacks under IHL are limited to those which target objects “whose total or partial destruction, capture or neutralization, in the circumstances ruling at the time [of the attack], offers a definite military advantage.”27 Intelligence is needed to affirm the necessity of a particular aerial strike on the basis of the nature of a target or its military importance to the adversary.28

6. The Expected incidental harm to civilians and civilian property. As the principle of proportionality dictates, any expected civilian harms must not be excessive in relation to the military advantage anticipated.29 To conduct this balancing act between harms and advantages, military commanders must be presented with, as accurate and as genuine as possible, assessments of the potential collateral damage and potential military advantage to ensue from an

&"&$ !""=$ "J2J$ :0".+1$ N'<+.0=$ )1001"$ .($ :*+A,00'<=$ 91&&'F1$ (".A$ 0;1$ 2"1&,51-0$

N"'-&A,00,-F$ 0;1$ PTQ[$ D'F*1$ C*<0*"'<$ 2".J1"08$ C.-71-0,.-# O9'8$ PR=$ PTTZU=$333#FJ.#F.7W(5&8&WJ@FWC!iCYPS_05./PW;0A<WC!iCYPS_05./P#;0A$Od!*",-F$iJ1"'0,.-$51&1"0$:0."A=$(."$1m'AJ<1=$,-01<<,F1-/1$"1&.*"/1&$31"1$*0,<,X15$0.$<..@$(."$/*<0*"'<$J".J1"08$,-$."51"$0.$J".J1"<8$,51-0,(8$,0#$N'"F10$,-01<<,F1-/1$.((,/1"&$,51-0,(,15$/*<0*"'<$J".J1"08$."$/*<0*"'<$J".J1"08$&,01&$,-$4"'ce$'$p-.Y&0",@1n$0'"F10$<,&0$3'&$J"1J'"15=$J<'/,-F$@-.3-$/*<0*"'<$J".J1"08$.(($<,A,0&$(".A$'00'/@=$'&$31<<$'&$&.A1$.0;1"3,&1$<1F,0,A'01$0'"F10&$,($0;1,"$'00'/@$A,F;0$J'</1$-1'"+8$/*<0*"'<$J".J1"08$'0$",&@$.($5'A'F1#fU#$

"'$!""#%24=$ %=?&'#-.01$PQ=#%"0#$QR#$>."$ (*"0;1"$5,&/*&&,.-$.($ 0;1$ <'3$ &*"".*-5,-F$'&&1&&A1-0&$.($A,<,0'"8$'57'-0'F1$&11$4CKC$kmJ1"0$9110,-F$K1J."0=$%=?&'#-.01$RR=$'0$PPYR\#$$

"($4-$9'8$RSPQ=$0;1$4&"'1<,$F.71"-A1-0$J*+<,&;15$'$"1J."0$&*AA'",X,-F$,0&$J.&,0,.-&$/.-/1"-,-F$0;1$171-0&$0;'0$*-(.<515$,-$0;1$&*AA1"$.($RSP[$5*",-F$.J1"'0,.-$d2".01/0,71$k5F1f#$N;1$"1J."0$3'&$'$"1&J.-&1$0.$0;1$,-01"-'0,.-'<$/.AA*-,08n&$/",0,/,&A$.($4&"'1<n&$'1",'<$ /'AJ',F-$ ,-$ M'X'$ 0;'0$ &*AA1"#$ N;1$ "1J."0$ J".7,51&$ 7,0'<$ ,-&,F;0$ ,-0.$ 0;1$ <1F'<$J.&,0,.-&$.($4&"'1<$"1<'0,-F$0.$'1",'<$0'"F10,-F$*-51"$4D)=$'-5$,-$A'-8$"1&J1/0&$,0$"1(<1/0&$0;1$J.&,0,.-&$.($.0;1"$/.*-0",1&$,-7.<715$,-$&,A,<'"$'1",'<$0'"F10,-F$.J1"'0,.-&$5*",-F$'/0,71$'"A15$/.-(<,/0#$!""#N-#&RSP[&M!%!&C3*4'+/$&O`&I('6&o&R_&%(,(.$&RSP[UV&>!/$(!'&!*2&

)#,!'& %.0#/$.=$ 4."!#'+& 9+*+.$"6& 34& >3"#+,*& %44!+".=$ q$ \PT$ O9'8$ RSPQU=$;00J&VWWA('#F.7#,<W2".01/0,71k5F1W2'F1&W51('*<0#'&Jm$ g;1"1,-'(01"V$ 4&"'1<$ RSP[$ M'X'$C.-(<,/0$ K1J."0h$ O&J1'@,-F$ 5,"1/0<8$ .-$ 0;1$ ,&&*1$ .($ ,-01<<,F1-/1$ (."$ '&&1&&,-F$ A,<,0'"8$'57'-0'F1=$ 0;1$ "1J."0$ -.01&V$ dg0hhis intelligence may include, for example, detailed information about the number and rank of militants anticipated to be hit during an attack, as well as the quality and quantity of enemy weapons expected to be destroyed. In performing proportionality analyses, commanders must focus on the concrete and direct military advantage anticipated and may not take into account unlikely possibilities of military advantage.fU#$$

")$%24=$ %=?&'#-.01$ PQ=$%"0#$ Q`O'UO,,,U$ Od"1("',-$ (".A$51/,5,-F$ 0.$ <'*-/;$ '-8$ '00'/@$3;,/;$A'8$+1$1mJ1/015$0.$/'*&1$,-/,51-0'<$<.&&$.($/,7,<,'-$<,(1=$,-G*"8$0.$/,7,<,'-&=$5'A'F1$0.$/,7,<,'-&$.+G1/0&=$."$'$/.A+,-'0,.-$0;1"1.(=$3;,/;$3.*<5$+1$1m/1&&,71$,-$"1<'0,.-$0.$0;1$/.-/"101$'-5$5,"1/0$A,<,0'"8$'57'-0'F1$'-0,/,J'015fU#$

A>& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

attack. Militaries rely on UAVs and informants to monitor civilian presence in the vicinity of the target as well as engineers and other technical experts who assist during the planning of the attack in determining their potential consequences amongst alternative means and methods of operation.30

Much of the above detailed intelligence will be collected and collated long before the operations are launched and often during peacetime. This information will be stored in military archives in order to ensure operational readiness and in preparation for a prospective war.31 To guarantee that the

&#*$ 4&"'1<$ RSP[$ M'X'$ C.-(<,/0$ K1J."0=$ %=?&'# -.01$ RZ=$ '0$ q$ \R_$ O-.0,-F$ 0;'0V$ dg0h;1$

1&0,A'0,.-$.($J.01-0,'<$/.<<'01"'<$5'A'F1$/'-$+1$71"8$/;'<<1-F,-F#$L.$A,<,0'"8$;'&$J1"(1/0$,-(."A'0,.-$"1F'"5,-F$0;1$J"1&1-/1$.($/,7,<,'-&$,-$'<<$0;1$'"1'&$3;1"1$'00'/@&$0'@1$J<'/1#$N;,&$,&$'<<$0;1$A."1$&.$3;1-$.J1"'0,-F$,-$'$/.AJ<1m$*"+'-$1-7,".-A1-0=$3,0;$51-&1$J;8&,/'<$,-("'&0"*/0*"1$'-5$'$A.+,<1$/,7,<,'-$J.J*<'0,.-#$b;,<1$A,<,0'",1&$'"1$"1c*,"15$0.$1m1"/,&1$5*1$5,<,F1-/1$'-5$0.$517.01$"1'&.-'+<1$1((."0&$0.$/.<<1/0$ ,-(."A'0,.-$3,0;$"1&J1/0$0.$0;1$/.<<'01"'<$ 5'A'F1$ 1mJ1/015=$ ,-(."A'0,.-$ 51(,/,0&$ '"1$ ,-17,0'+<1#fU#$ 4$ 'F"11$ 3,0;$ 0;,&$&0'01A1-0#$D.3171"=$,0$,&$J"1/,&1<8$0;1$-."A'0,71$+<'/@$+.m1&$0;'0$'"1$"1(1""15$0.$;1"1$'&$d-="#-)()2"1A"f=$d&"'%/1'9("#"FF/&.%f=$'-5$d,-(."A'0,.-$51(,/,0&f$0;'0$4$3'-0$0.$.J1-$*J$,-$0;,&$%"0,/<1$'-5$1mJ<."1$(*"0;1"#$!""#'(%/$4CKC$kmJ1"0$9110,-F$K1J."0=$%=?&'#-.01$RR=$'0$PZ$O0;1$1mJ1"0&$d1AJ;'&,X15$0;'0=$,-$'"A15$/.-(<,/0&=$0;1$51F"11$.($/1"0',-08$.($'/;,17,-F$0;1$ '-0,/,J'015$ A,<,0'"8$ '57'-0'F1$ ,&$ *-<,@1<8$ 0.$ +1$ PSS$ J1"$ /1-0=$ 5*1$ 0.$ ,-&*((,/,1-0$,-(."A'0,.-=$J."01-0,'<$1-1A8$/.*-01"YA1'&*"1&$."$0;1$(.F$.($3'"#$:.A1$1mJ1"0&$1mJ"1&&15$0;1$7,13$0;'0$0;1$"1<17'-0$A,<,Y$0'"8$'57'-0'F1$0;1"1(."1$51J1-515$.-$0;1$/.A+,-'0,.-$.($0;1$A'F-,0*51$.($0;1$'-0,/,J'015$A,<,0'"8$'57'-Y$0'F1$'-5$0;1$<,@1<,;..5$.($.+0',-,-F$,0#$2*0$.0;1"3,&1=$0;1$*-/1"0',-08$.($.+0',-,-F$0;1$'-0,/,J'015$A,<,0'"8$'57'-0'F1$5,A,-,&;1&$0;1$/.-/"101-1&&$.($0;1$'57'-0'F1$0.$+1$0'@1-$,-0.$/.-&,51"'0,.-#$>."$1m'AJ<1=$,($'-$'00'/@$,&$<'*-/;15$'F',-&0$ '$A,<,0'"8$.+G1/0,71=$ '-5$ 0;1"1$ ,&$ '$QS$J1"$ /1-0$ /;'-/1$ 0;'0$ '-$1-1A8$/.AA'-51"$3,<<$+1$J"1&1-0$,-$0;1$+*,<5,-F$'0$0;1$0,A1$.($0;1$&0",@1=$0;,&$J.&&,+,<,08$3.*<5$,-/"1'&1$0;1$A,<,0'"8$'57'-0'F1$'-0,/,J'015$O'-5$0;*&$0;1$A'F-,0*51$.($,-/,51-0'<$/,7,<,'-$;'"A$0;'0$3.*<5$-.0$+1$1m/1&&,71$,-$"1<'0,.-$0.$,0U=$+*0$0.$'$<1&&1"$1m01-0$0;'-$,($0;1$J"1&1-/1$.($0;,&$/.AA'-51"$3'&$/.-(,"A15#fU#$

#!$4-$4&"'1<,$A,<,0'"8$G'"F.-=$(."$1m'AJ<1=$0;1&1$'"/;,71&$'"1$.(01-$"1(1""15$0.$'&$d0'"F10$+'-@&f=$&11$,-01"7,13$3,0;$9'G."$:#=$0;1$51J*08$/.AA'-51"$.($0;1$4&"'1<,$%,"$>."/1$RSS$:c*'5".-=$&"?&)1."-#)1#%--$K.F1"&=$4-71&0,F'0,-F$0;1$K1<'0,.-&;,J$+10311-$!".-1$b'"('"1$'-5$C,7,<,'-$C'&*'<0,1&$,-$M'X'=$`$I#$:0"'01F,/$:1/#$T[=$PSP$ORSP[U$g;1"1,-'(01"V$4-01"7,13$3,0;$9'G."$:#h$Od%0$'-8$F,71-$A.A1-0=$0;1$&c*'5".-$;'&$6%E&$,-$0;1$',"###$b1$F'0;1"$'$<.0$.($,-(."A'0,.-$3;,/;$171-0*'<<8$F,71&$*&$0;1$'+,<,08$0.$5101/0$ 0'"F10&$ 0;'0$-115$0.$+1$'00'/@15#$N;,&$,&$3;8$0;1$A,-*01$iJ1"'0,.-$2".01/0,71$k5F1$+1F'-=$0;1$',"$(."/1$'<"1'58$;'5$'$<'"F1$p+'-@n$.($0'"F10&#fU#$b,0;,-$0;1$+'-@$1'/;$0'"F10&$"1/1,71&$'$51&,F-'015$d/'"5f=$&11$4&"'1<$RSP[$M'X'$C.-(<,/0$K1J."0=$%=?&'#-.01$RZ=$'0$q$R[_$O(.<<.3,-F$'$7,&*'<,X'0,.-$.($'-$'/0*'<$ &*/;$/'"5$ (".A$ 0;1$+'-@=$ 0;1$ "1J."0$J"./115&$ 0.$ /<'",(8V$d0;1$ 0'"F10$J<'--,-F$J"./1&&$+1F,-&$3,0;$0;1$/.<<1/0,.-$.($,-01<<,F1-/1$,-(."A'0,.-$"1F'"5,-F$0;1$J.01-0,'<$0'"F10$'&$31<<$,&$,0&$&*"".*-5,-F&=$"1(<1/015$g###h$,-$0;1$<1(0$;'-5$/.<*A-$.($0;1$N'"F10$C'"5$0,0<15$d4-01<<,F1-/1#f$6&,-F$0;,&$,-(."A'0,.-=$/.AA'-51"&$5101"A,-1$0;1$.+G1/0,71&$3,0;$"1F'"5$0.$0;1$0'"F10=$,-/<*515$,-$0;1$N'"F10$C'"5$*-51"$0;1$+<./@$0,0<15$diJ1"'0,.-'<$!,"1/0,71&f###$i-$ 0;1$ +'&,&$ .($ 0;1$ .+G1/0,71&$ ,-$ '00'/@$ '-5$ 0;1$ ,-01<<,F1-/1$ ,-(."A'0,.-=$ .J1"'0,.-'<$J<'--1"&$ '&&1&&$ 0;1$ .J0,.-'<$ A1'-&$ '-5$ A10;.5&$ (."$ /'""8,-F$ .*0$ '$ J.01-0,'<$ '00'/@=$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& AD&

information remains accurate, it will be routinely inspected, reviewed, and updated on the basis of new intelligence.32 As the armed conflict continues to progress the preplanned targets will begin to dwindle and militaries will move to engage in interdicting enemy forces discovered in real time. This happened in Afghanistan where the Joint Surveillance and Target Attack Radar System (JSTARS) and UAVs of the U.S. air force were relied upon together with information from on-ground Special Operations Forces to spot “pop-up targets” and relay “time-sensitive up-to-date target information to shooter platforms that were either inbound or already circling in the vicinity.”33 In the age of Artificial Intelligence, such real-time intelligence collection, analysis, and dissemination is only likely to increase and get further automated.34

&,-/<*5,-F$'&&1&&,-F$3;'0$J"1/'*0,.-&$/.*<5$+1$0'@1-$5*",-F$0;1$'00'/@$J"./1&&e$0;1&1$.J0,.-&$'"1$,-/<*515$*-51"$0;1$+<./@$0,0<15$diJ1"'0,.-'<$2<'--,-Ff###$%&$J'"0$.($0;1$J"./1&&=$<1F'<$'57,&1"&$J".7,51$0;1,"$+,-5,-F$.J,-,.-$'&$0.$0;1$<1F'<,08$.($'-$'00'/@=$J<*&$'-8$&0,J*<'0,.-&$(."$,0&$1m1/*0,.-$O,-/<*5,-F$/.-/1"-,-F$J"1/'*0,.-&$,-$'00'/@Ue$0;,&$'JJ1'"&$,-$0;1$",F;0$.($0;1$N'"F10$C'"5=$*-51"$0;1$0,0<1$d)1F'<$K17,13#f$?18.-5$0;1$<1F'<$,-J*0=$/.AA'-51"&$A'8$,-&1"0$'55,0,.-'<$/.-5,0,.-&=$3;,/;$'"1$,-/<*515$,-$0;1$diJ1"'0,.-'<$!,"1/0,71&f$+<./@###$N;1$A'--1"$,-$3;,/;$0;1$N'"F10$C'"5$,&$&0"*/0*"15=$0;1$1-0,0,1&$,-7.<715=$'-5$0;1$J"./1&&$"1c*,"15$ ,-$ (,<<,-F$ .*0$ &*/;$ '$ 5./*A1-0=$ "1(<1/0&$ 0;1$ ,-01F"'0,.-$ .($ 0;1$ )'3$ .($%"A15$C.-(<,/0$3,0;,-$0;1$4!>n&$0'"F10,-F$J"./15*"1&#fU#$

#"$!""#4-01"7,13$3,0;$9'G."$:#=$%=?&'#-.01$\P=$'0$PS\YPS[$Odg1h71"8$(13$A.-0;&=$,0$,&$1&&1-0,'<$ 0.$/;1/@$ 0;'0$ 0;1$ 0'"F10$ ,&$&0,<<$ "1<17'-0#$ 4($8.*$(,-5$'$31'J.-&$&0."'F1$('/,<,08$0.5'8=$ 0.A."".3$ 0;18$ /.*<5$ 0'@1$ '<<$ .($ 0;1$ 31'J.-&$ .*0$ .($ 0;1$ +*,<5,-F$ '-5$ +*,<5$ '$@,-51"F'"01-#$4($4$5.-n0$@-.3$'+.*0$0;'0$/;'-F1=$4$A,F;0$'//,51-0'<<8$0'"F10$,0#$N;'0n&$3;8$31$5.-n0$.-<8$(,-5$-13$0'"F10&e$31$'<&.$@11J$0"'/@$.($0;1$1m,&0,-F$.-1&#fU#$

##$!""#i&,-F'$a$9'"@$K.."5'=$%=?&'#-.01$RP=$'0$_S#$#$$ The Joint Artificial Intelligence Center (JAIC) sits at the cornerstone of the

Department of Defense (DoD)’s efforts to deliver Artificial Intelligence (AI) and other algorithm-based decision-making solutions to warfighters. The JAIC’s budget grew from $89 million in 2019 to a whopping $268 in 2020, with the DoD intending to spend close to $1.6 billion over the next few years on the evolution of wartime AI applications. These applications seek to offer real time support for front-line commanders which is “critical in an evolving operational environment where speed, precision, and agility are paramount for success.” (!""#!'7,5$E1"F*-=$W"()L"&)12#*$# ./#;'&F)24."&%# $%#!.&'."2)A# $+?"&'.)L"=$6:&

!#0n$& !#4#$ O:1J0#$ PS=$ RSRSU=$;00J&VWW333#51(1-&1#F.7WkmJ<."1WL13&W%"0,/<1W%"0,/<1WR\[\QSSW51<,71",-FY',Y0.Y3'"(,F;01"&Y,&Y&0"'01F,/Y,AJ1"'0,71WUe$ Indeed, AI military technologies are constantly evolving and improving and may prove useful in an array of different operational environments: “N;1$J1"(."A'-/1$.($0;1&1$&8&01A&$/'-$A'@1$0;1A$71"8$*&1(*<$(."$0'&@&$&*/;$'&$,51-0,(8,-F$'$NYTS$A',-$+'00<1$0'-@$,-$'$&'01<<,01$,A'F1=$,51-0,(8,-F$;,F;Y7'<*1$0'"F10&$,-$'$ /".35$*&,-F$ ('/,'<$ "1/.F-,0,.-=$ 0"'-&<'0,-F$ 01m0$ (."$ .J1-Y&.*"/1$ ,-01<<,F1-/1=$ '-5$ 01m0$F1-1"'0,.-$ (."$ *&1$ ,-$ ,-(."A'0,.-$ .J1"'0,.-&#$ %4$ ,&$ '<&.$ 71"8$ /'J'+<1$ ,-$ '"1'&$ <,@1$"1/.AA1-5'0,.-$&8&01A&=$'-.A'<8$5101/0,.-=$J"15,/0,.-$&8&01A&=$'-5$/.AJ10,0,71$F'A1&#$%-$ %4$ &8&01A$ ,-$ 0;1&1$ 5.A',-&$ /.*<5$ '&&,&0$ 0;1$ A,<,0'"8$ 3,0;$ ("'*5$ 5101/0,.-$ ,-$ ,0&$/.-0"'/0,-F$&1"7,/1&=$J"15,/0,-F$3;1-$31'J.-&$&8&01A&$3,<<$(',<$5*1$0.$A',-01-'-/1$,&&*1&=$

AO& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

In identifying the normative framework surrounding wartime intelligence production we should distinguish between two separate legal questions: (1) where do we derive the right to spy in times of war, or what is the “Jus Ad Explorationem” (JAE); (2) what laws govern the conduct of intelligence collection, analysis and promulgation, or what is the “Jus In Exploratione” (JIE). These two questions correlate with two temporal stages in the practice of intelligence operations: before launching the operation and during the operation. There is a third final temporal stage, relating to the ramifications of wartime spying, once the spying, and likely the war, have ceased. This is the “Jus Post Explorationem” (JPE).35 This section will review the law surrounding the first two temporal stages (JAE, JIE). I make general observations relating to the third stage (JPE) in section III of this paper.

+/&&012&*3415&56&.78&39&:;<&

Should wartime spying be prohibited as a violation of the international legal principles of sovereign equality, territorial integrity, and non-intervention? The late Dr. Kish denied the existence of such a prohibition, suggesting instead that as “war prevails over sovereignty, the peacetime prohibition of espionage in national spaces does not apply between belligerent States, and the permissibility of espionage extends to all areas of hostilities, land, sea, and air, whether national or international.”36 This is quite intuitive. If war itself is legal, that is to say if a State may lawfully launch armed attacks and acts of belligerency against another under certain constraints laid down in the U.N. Charter, then certainly non-armed forms of intervention (such as intelligence gathering) targeted against the enemy State, and serving the same war efforts, must also be lawful.37

&."$5171<.J,-F$3,--,-F$&0"'01F,1&$,-$/.-(<,/0$&,A*<'0,.-&#$%<<$.($0;1&1$'JJ<,/'0,.-&$'-5$A."1$/'-$+1$(."/1$A*<0,J<,1"&$,-$5'8Y0.Y5'8$.J1"'0,.-&$'-5$,-$0;1$-1m0$/.-(<,/0#” (2'*<$9'm31<<=$*&.)F)A)'(#$1."(()2"1A"#)%#.4"#7=.=&"#/F#;'&F'&"#XN=%.#Y/.#$1#<4"#;',#Z/=#<4)1[\=$932#"*&

b!"& 4*.$+$($#$ O%J"#$ [=$ RSRSU=$ ;00J&VWWA3,#*&A'#15*W'"0,(,/,'<Y,-01<<,F1-/1Y(*0*"1Y3'"('"1YG*&0Y-.0Y3'8Y0;,-@W).$

#%#4$1<'+."'01$(*"0;1"$.-$0;,&$0;"11Y(.<5$("'A13."@$F.71"-,-F$0;1$<'3$.($,-01<<,F1-/1$,-$1'"<,1"$3."@#$:11$)*+,-=$%=?&'#-.01$PT=$'0$RPSYRPP#$

36 B,&;, supra note 16, at 123. 37 The customary principle of non-intervention forbids States from intervening directly

or indirectly in the internal affairs of other States (See, e.g., Declaration on the Principles of International Law Concerning Friendly Relations and Cooperation among States GA Res 2625 (XXV) (1970), Principle 3). In the Nicaragua Case, this International Court of Justice identified coercion that undermines sovereign free choice as forming the very essence of intervention (Military and Paramilitary Activities in and Against Nicaragua (Merits) (Nicar. v. U.S.), 1986 I.C.J. 14 (June 27), q205). “Coercion" is not limited to the use of force, but includes "all other forms of interference or attempted threats", thus precluding States from using economic, political or any other type of measures to coerce another state in order to

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& AW&

The permissibility of wartime intelligence collection is also apparent from reading the treatises of IHL. Article 101 to the Lieber Code of 1863 provided an early codification of the basic principle that “deception in war is admitted as a just and necessary means of hostility and is consistent with honourable warfare.”38 Article 13(e) of the Russian Draft Convention on the Laws and Customs of War presented at the 1874 Brussels International Conference followed a similar route suggesting that: “amongst the means of warfare which are permitted are the employment of every available means of procuring information about the enemy and the country.”39 These positions were solidified in Article 24 of both the 1899 and the 1907 Hague Regulations, which are also reflective of customary international law. The latter Convention reads: “Ruses of war and the employment of measures necessary for obtaining information about the enemy and the country are considered permissible.”40 These sources all confirm what was concluded by the great Richard Baxter as early as 1951: “espionage is regarded a conventional weapon of war, being neither treacherous nor productive of unnecessary suffering.”41 In the context of intelligence collection for aerial

&obtain from it the subordination of the exercise of its sovereign rights or to secure from it advantages of any kind (Declaration on the Inadmissibility of Intervention on the Domestic Affairs of States and the Protection of their Independence and Sovereignty, G.A. Res. 2131, U.N. GAOR, 20th Sess., Supp. No 14, Dec. 21, 1965). However, at war this framework ceases to exist. Countries at war revert to arms precisely because they wish to undermine sovereign free choice and subordinate the exercise of another sovereign’s rights.

38 Francies Lieber, Instructions for the Government of Armies of the United States in the Field, Article 101, originally published as U.S. WAR DEPARTMENT, ADJUTANT GENERAL’S OFFICE, GENERAL ORDERS NO. 100 (Apr. 24, 1863).

39 Russian Draft Convention on the Laws and Customs of War, Article 13(e) (July 27, 1874), reprinted in 2 DIETRICH SCHINDLER AND JIŘÍ TOMAN, THE LAWS OF ARMED CONFLICTS: A COLLECTION OF CONVENTIONS, RESOLUTIONS, AND OTHER DOCUMENTS 29 (1988).

40 See Regulations Respecting the Laws and Customs of War on Land, Annex to Convention (IV) Respecting the Laws and Customs of War on Land, 36 Stat. 2295, Article 24 (Oct. 18, 1907) [hereinafter: Hague Regulations 1907]; See also Regulations Respecting the Laws and Customs of War on Land, annexed to Convention with Respect to the Laws and Customs of War on Land, 32 Stat. 1803, 1811, Article 24 (Jul. 29, 1899).

41 Richard R. Baxter, So-Called ‘Unprivileged Belligerency’: Spies, Guerillas, and Saboteurs, 28 Brit. Y.B. Int’l L. 323, 333. See also, United States v. List, et al. (The Hostages Case), XI TRIALS OF WAR CRIMINALS BEFORE THE MNT 1245 (“By the law of war it is lawful to use spies”). As the Department of Defense Law of War Manual notes: “Spying [is] not prohibited by any law of war treaty to which the United States is a party. For example, spying [is] not prohibited by the 1949 Geneva Conventions, nor defined as a “grave breach” of those conventions. Similarly, spying [also has] not been listed as war crimes punishable under the statutes of international criminal tribunals. In addition, law of war treaties that regulate, but do not prohibit, spying, recognize implicitly that belligerents may use this method of warfare.” (see, DEPARTMENT OF DEFENSE LAW OF WAR MANUAL, OFFICE OF GENERAL COUNSEL, q$ [#P`#\ (June 12, 2015)

AQ& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

targeting specifically the justification was further clarified in Jean Pictet’s commentary to the First Additional Protocol to the Geneva Conventions (API):

“In the case of long-distance attacks, information will be obtained in particular from aerial reconnaissance and from intelligence units, which will of course attempt to gather information about enemy military objectives by various means.”42

Foreign Intelligence gathering was thus perceived by Pictet not only as a lawful measure during war, but rather as a prerequisite for the proper functioning of military units in compliance with their other IHL obligations. The only way to comply with the rules of distinction, military necessity, and proportionality, is by an unremitting wartime intelligence campaign.43

B. 012&!;=&>6?2<9394&$952@@3429A2&)72<;5369B&39&:;<53C2 Before we begin analyzing the law that applies to the conduct of wartime

intelligence operations it is important to briefly explain the structure of a typical process of intelligence production. What is commonly referred to as the “intelligence cycle” is better understood as “the steps or stages in intelligence, from policy makers perceiving a need for information to the community’s delivery of an analytical intelligence product to them.”44 Professor Lowenthal maps out seven such steps common to most intelligence

&www.dod.mil/dodgc/images/law_war_manual15.pdf (hereinafter: DOD Manual).

42 Jean S. Pictet et al., Commentary on the Additional Protocols of 8 June 1977 to the Geneva Conventions of 12 August 1949, at 681 q$2195 (International Committee of the Red Cross, 1987).

43 Without delving too deeply into the subject, we should recognize the difficulty with this general line of justification under Jus Ad Explorationem. As already discussed, a vast majority of intelligence collection for aerial targeting purposes is done in peacetime, in preparation for the war. This brings to the forefront questions concerning the legality of peacetime espionage, an issue left unresolved in the literature. Indeed, the U.N. Charter principles of territorial integrity and non-intervention, those abrogated during armed conflict, are very much in effect in times of peace. The extent to which we can use a wartime privilege to justify potential violations of international law during peacetime, when the war is only looming, is highly contentious (for further discussion see Lubin, supra note 19, at 224-25). Moreover, if we derive the legality of peacetime intelligence collection activities from their later wartime use, we risk conflating Jus Ad Bellum with Jus In Bello; two legal regimes which should operate independently of one another (for a broader discussion on the relationship between Jus Ad Bellum and Jus In Bello see DOD Manual, supra note 36, at 86-89, q$\#Q#P). Nonetheless, for the purposes of this paper it is enough to establish undisputedly that, to the very least, wartime intelligence gathering is permissible as a matter of customary IHL.

44 LOWENTHAL, INTELLIGENCE: FROM SECRETS TO POLICY, 70 (6th ed., 2014).

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& A[&

community cycles: (1) identification of requirements; (2) collection of intelligence information; (3) processing and exploitation of said information;45 (4) collation, analysis and production of intelligence products; (5) dissemination of products; (6) consumption of products by policy makers; and (7) feedback, which leads to identification of new requirements, and the wheel goes round.46 To these seven stages, I would add a continuous eighth stage: (8) verification of authenticity and accuracy. Throughout the entire cycle, and parallel to the other tasks of collection, processing, analysis, consumption, and feedback, materials must be routinely scrutinized. The credibility of the information must be closely examined and consistently challenged.

How do these eight steps translate into the language of IHL? Article 24 of the Hague Regulations is worth repeating. In times of war “the employment of measures necessary for obtaining information about the enemy and the country are considered permissible.”47 As the commentary provides this provision should not be taken to mean that “every ruse of war and every method necessary to obtain information about the enemy and the country should ipso facto be considered permissible.”48 Indeed “in obtaining information, a belligerent must not contravene specific rules of war.”49 As the U.S. Department of Defense’s (DoD) law of war manual highlights: “it would be unlawful, of course, to use torture or abuse to interrogate detainees for purposes of gathering information. Similarly, it is prohibited to make improper use of a flag or truce to obtain information.”50

In both the treatises and customary law of IHL one can find a long list of prohibitive and clarifying rules covering an ambit of possible intelligence collection activities. Aside from the two examples already cited in the quote from the DoD’s manual, one can, inter alia, find answers in IHL to additional questions such as: Whether collecting VISINT from a balloon or other military reconnaissance aircrafts constitutes a legitimate act of war?51 Whether a neutral vessel may collect intelligence on the activities of a belligerent and transmit this information to another belligerent, or whether a

&$%$N;1$J"./1&&$.($51/.5,-F$'-5$"1."F'-,X,-F$0;1$,-(."A'0,.-$,-$'$3'8$0;'0$A'@1&$,0$

'//1&&,+<1$0.$0;1$'-'<8&0&#$N;1$A.&0$/.AA.-$'-5$&,AJ<1$1m'AJ<1$(."$'/0,.-$.($0;,&$&0'F1$,&$0"'-&<'0,-F$5./*A1-0&$(".A$0;1,"$.",F,-'<$<'-F*'F1$0.$0;1$&J.@1-$<'-F*'F1$.($0;1$'F1-/8#$$

46 Lowenthal, supra note 44, at 70. $'$!""#%=?&'#-.01$[S$'-5$'//.AJ'-8,-F$01m0#$$($K1J."0$.($0;1$:1/.-5$C.AA,&&,.-=$C.AA1-0'"8$0.$!"'(0$%"0,/<1$R[=$4-01"-'0,.-'<$

21'/1$C.-(1"1-/1=$N;1$D'F*1=$Q$I*<8$PZTT=$&"?&)1."-#)1#I!)#.&?#&:/3$$=&K#03"$.&$3&

$-#&D!,(#&C3*4#"#*/#.&34&PZTT&!*2&PTS`=$P[_$OPTP`U#$$)$93""+.&M"##*.0!*=&N-#&932#"*&)!5&34&)!*2&b!"4!"#$\RQ$OPT_SU#$%*$See DOD Manual, supra note 41, at q$Q#R_#R#$%!$D'F*1$K1F*<'0,.-&$PTS`=$%=?&'#-.01$[S=$'0$%"0,/<1$RT#$

Ad& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

belligerent may conduct espionage from neutral ports and waters?52 Whether propaganda may be used to support intelligence gathering?53 Whether POWs, protected persons, and the civilian population in the occupied territory, may be subjected to pressure in order to obtain certain intelligence information?54 Whether belligerents may continue to gather intelligence during an armistice, and may they rely on that intelligence after the armistice had ended?55 The list goes on.

What is important to note, however, is that all of these rules relate only to the second stage of the intelligence cycle. While there exists ample regulation as to the way intelligence is to be collected during war, the other stages of the intelligence cycle, in particular intelligence processing and exploitation, analysis, and verification (3, 4, and 8 respectively) remain unregulated. This is problematic as it is in these stages where the bleeding gets done in IHL, quite literally.

There is only one rule in the treatises from which one can derive some general guideline on how intelligence is to be processed, analyzed, and verified in times of war. The rule is enumerated under Article 57(2)(a)(i)-(ii) to API:

“Those who plan or decide upon an attack shall: (i) do everything feasible to verify that the objectives to be attacked are neither civilians nor civilian objects and are not subject to special protection but are military objectives within the meaning of paragraph 2 of Article 52 and that it is not prohibited by the provisions of this Protocol to attack them; (ii) take all feasible precautions in the choice of means and methods of attack with a view to avoiding, and in any event to minimizing, incidental loss of civilian life, injury to civilians and damage to civilian objects”.56 [highlights added – AL]

How should we interpret the phrase “do everything feasible to verify” or

“take all feasible precautions” as a matter of a legal standard? The Expert Committee established by the prosecutor of the International Criminal Tribunal for the Former Yugoslavia (ICTY) to review the 1999 NATO bombing campaign, attempted to answer this question:

&%"$!""$B+.-, supra note 16, at 128-133#$%#$See DOD Manual, supra note 41, at q$Q#R_#P#$%$$!""$B+.-, supra note 16, at 137-144#$%%$See DOD Manual, supra note 41, at q$PR#PP#[#[#$%&$ %24=$ %=?&'# -.01$ PQ=$ '0$ %"0,/<1$ Q`ORUO'UO,UYO,,U#$ !""# '(%/# 4CKC=& C(.$3)!"6&

4*$#"*!$+3*!'& D()!*+$!"+!*& )!5=$ K*<1$ P_$ O7.<#$ 4=$ RSSQU$ g;1"1,-'(01"$ 4CKC$C*&0.A'"8$K*<1&h#$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& AJ&

“The obligation to do everything feasible is high but not absolute. A military commander must set up an effective intelligence gathering system to collect and evaluate information concerning potential targets. The commander must also direct his forces to use available technical means to properly identify targets during operations. Both the commander and the aircrew actually engaged in operations must have some range of discretion to determine which available resources shall be used and how they shall be used”.57 [highlights added – AL]

While the Committee’s clarification is certainly a step forward, in the

sense that it tries to elaborate on the obligations of the contracting parties, as it stands it is quite problematic. The first problem with the Committee’s approach is that it replaces one ambiguous phrase (do everything feasible) with two others (effective intelligence gathering system, properly identify). Moreover, following the language of API Article 57 the Committee’s report vests all responsibility with the military commander. She is required “to set up an effective intelligence gathering system”. She must “direct [her] forces to use available technical means to properly identify targets”. She should have the “discretion to determine which available resources shall be used”.

This focus on the commander comports with traditional thought. Tribunals and military manuals guide us to rely on the “reasonable commander” test in determining the lawfulness of a particular strike.58 Yet, in the process we neglect the fact that any reasonable commander will turn to her “reasonable intelligence analyst,” the contours of this standard are notably under-defined. Modern military commanders often do not know what “an effective intelligence gathering system” looks like, or how to properly process, analyze, and verify intelligence materials as a matter of professional practice. Nor should they be expected to do so––it would be inefficient as that

&%'$>,-'<$K1J."0$0.$0;1$2".&1/*0."$+8$0;1$C.AA,0011$k&0'+<,&;15$0.$K17,13$0;1$L%Ni$

?.A+,-F$C'AJ',F-$%F',-&0$ 0;1$ >151"'<$ K1J*+<,/$ .($H*F.&<'7,'=$ q$ RT$ OI*-1$ R=$ RSSSU$;00J&VWW333#,/08#."FW&,5WPSSQR$O;1"1,-'(01"V$4CNH$kmJ1"0$C.AA,0011$K1J."0U#$$

%($See ICRC, CUSTOMARY INTERNATIONAL HUMANITARIAN LAW: PRACTICE – PART 1 331-335 (vol. II, 2005); Prosecutor v. Stanislav Galić, Case No. IT-98-29-T, Judgment and Opinion, Trial Chamber I, para. 58 (2003) ("In determining whether an attack was proportionate it is necessary to examine whether a reasonable well-informed person in the circumstances of the actual perpetrator, making reasonable use of the information available to him or her, could have expected excessive civilian casualties to result from the attack"); H.C.J. 769/02, The Public Committee Against Torture in Israel v. Government of Israel (Dec. 11, 2005), q$ 57 (Isr.) http://www.haguejusticeportal.net/Docs/NLP/Israel/Targetted_Killings_Supreme_Court_13-12-2006.pdf [hereinafter: Targeted Killings Case] (“the question is not what I would decide in the given circumstances, rather whether the decision which the military commander made is a decision that a reasonable commander was permitted to make.”).$

>a& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

is outside the scope of their professional expertise. To exemplify the problem, consider the Obama-era Presidential Policy

Guidance on drone warfare. In 2013 the administration introduced a policy whereby aerial strikes are to be authorized only when “there is near certainty that the individual being targeted is in fact the lawful target and located at the place where the action will occur.”59 Also, the policy established that such strikes may be launched only if there is “near certainty that the action can be taken without injuring or killing non-combatants.”60

On its face, this is quite a promising guidance from the perspective of humanitarian protection. Nonetheless, once the guidance leaves the desk of the President, it must be operationalized by dozens of military, intelligence, and legal professionals. This is where the policy enters the great charcuterie of the intelligence community. The professionals take the words of the president, the recipe if you will, and begin cooking. The dish they serve might look quite different from what was originally envisaged, as the general guidance is churned and minced within the intelligence community’s many meat grinders and sausage makers. Determinations of “near certainty” depend on intelligence policies and assessment guidelines which are outside the scope of expertise of the commander-in-chief.61

The story of Warren Weinstein and Giovanni Lo Porto, an American and Italian al-Qaeda hostages killed in an American drone strike in January 2015, offers a good example of this. If an intelligence agency, after “hundreds of hours of surveillance,” authorizes an operation to target an al-Qaeda compound without any idea that the hostages were being held there, and then months later still doesn’t even realize that it had killed them, this is a gross intelligence omission.62 The responsibility, as do the expectations, should be vested with the intelligence agency and not solely with the commander-in-chief. Intelligence agencies should operate under clearer rules and guidelines as to the way their assessments are to be developed so to comply with the general obligation to take all feasible precautions as enshrined under API.

&%)$Procedures for Approving Direct Action Against Terrorist Targets Located Outside

the United States and Areas of Active Hostilities, 1 (May 22, 2013), http://www.nuhanovicfoundation.org/user/file/2016_presidential_policy_guidance.pdf. This Presidential Policy Guidance (PPG) was released in a redacted version in August 2016 under court order. $

&*$Id.$&!$The Presidential Policy Guidance itself called on “operating agencies” to establish

“harmonized policies and procedures for assessing: (1) near certainty that a lawful target is present; (2) near certainty that non-combatants will not be injured or killed.” In other words, the guidance specifically asks the chefs to figure out how to make the sausage. Id., at 4.$

&"$See Peter Baker, Obama Apologizes After Drone Kills American and Italian Held by Al-Qaeda, N.Y. TIMES (Apr. 23, 2015) https://www.nytimes.com/2015/04/24/world/asia/2-qaeda-hostages-were-accidentally-killed-in-us-raid-white-house-says.html?_r=0.$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& >A&

This is what is the ICTY’s Expert Committee Report intentionally left out. Justice Barak, in the 2005 Israeli Supreme Court’s Targeted Killing case,

one of the most cited cases in the context of the law surrounding aerial strikes, took a minor step in standardizing wartime intelligence production. He noted in his judgment that “well based information is needed” for launching an attack, further clarifying that such information must be “thoroughly verified” and that the “burden of proof on the attacking army is heavy.”63 While it is true that Barak has too fallen into the trap of fungible metaphors and turns of phrase, his latter point, suggesting an evidentiary standard for intelligence in targeting situations, is commendable. This evidentiary standard, however, must be further developed into practical guidelines for those intelligence professionals who are required to meet it. Otherwise, we will find ourselves with a growing chasm between policy and practice. As Jameel Jaffer, former deputy legal director of the American Civil Liberties Union said days after the revelation of the CIA strike that killed the two hostages: “[t]hese and other recent strikes in which civilians were killed make clear that there is a significant gap between the relatively stringent standards the government says it’s using and the standards that are actually being used.”64

It is precisely this gap that this paper seeks to fill in Section III below. My proposal attempts to open up the ‘black boxes’ of wartime intelligence standards, look inside, and offer some practical substance. This substance will be identified and derived by examining five different historical case studies of intelligence failures in aerial strikes, spanning geographical regions, military doctrinal backgrounds, and belligerent parties.

CC<& 8%"()L#.-/("9)R%.'#3)C,#(''/0(,&()/,)=($/%')L#$/4(")

The academic analysis of historical intelligence failures has been by far dominated by intelligence studies scholars, and to a lesser extent by researchers from political science and foreign affairs departments.65 Rarely have lawyers entered the ring. As early as 1984 Professor W. Michael Reisman identified the reasons for why international lawyers tend to pay “little attention to the incidents from which political advisers infer their normative universe.”66 Reisman suggested that international lawyers “persist

&&#$See Targeted Killings Case, supra note 58, at q$[S#$&$$See Baker, supra note 62.$&%$See, e.g., K,/;'"5$B#$?100&=$*1'(,%)%B#;'&B#'1-#W"A)%)/1:#;4,#$1."(()2"1A"#7')(=&"%#

'&"# $1"L).'9("=$ \POPU$ b3"'2& 23'#$ _P$ OPT`ZUe$ k",/$ K.&1-+'/;=$ <4"# $1A)%)L"# 7)24.:#C"A/++"1-'.)/1%# F/&# $+?&/L)12# K/=1."&."&&/&)%+# $1."(()2"1A"=$ _PZOPU$ %LL%):$ P\\$ORSSZUe$k",@$!';<=$8)%%)12#.4"#;'["R]?#K'((:#;4,#$1."(()2"1A"#7')(=&"%#C'&"(,#$1%?)&"#$+?&/L"-#O"&F/&+'1A"=$RQO_U$4*$#''+,#*/#&a&L!$n'&:#/#$``Z$ORSPSU#$

&&$W. Michael Reisman, The Incident as a Decisional Unit in International Law, 10(1)

>>& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

in constructing their normative universe from texts. They thus confine their attention to sources of international law that were either merely ceremonial at their inception, or that, although animated by more normative intentions when they were created, have ceased to be congruent with expectations of authority and control heled by effective elites.”67

ng a proposal for the further regulation of the intelligence function in IHL, it is crucial that we remember the lessons of previous attacks and review past examples of intelligence incidents in the conduct of aerial warfare. In so doing we will be able to derive from each incident “norm-indicators” and “norm-generators”68 which could serve the purpose of further elucidation of existing and future expectations and rules of the road. Such an incident analysis could serve as a type of what Reisman called “meta-law” providing “normative guidelines for decisionmakers in the international system in those vast deserts in which case law is sparse.”69

Reisman, in fact, did exactly that in his 1997 comment “The Lessons of Qana” which explored the law and politics of a 1996 Israeli artillery strike at a United Nations Compound in Southern Lebanon.70 The analysis I follow below adopts a similar methodology to that employed by Reisman in that article.

A. Case Selection Criteria

B0#&S/I#&%#"#.*#'&.$%#&%*,'/#%&$""&/)I("I#&S$,"*M&/)*#""/5#).#&:M&$&)$*/()3

%*$*#& #)5$5#'& /)& $-7#'& .()S"/.*<& B0#%#& .$%#& %*,'/#%& S,-*0#-& 7##*& *0#&S(""(1/)5&.-/*#-/$9&VAY&*0#&*$-5#*/)5&L*$*#&0$'&$'7/**#'&S$,"*X&V>Y&%,:%*$)*/I#&#I/'#).#&/%&$..#%%/:"#&*(&'#7()%*-$*#&*0#&/)*#""/5#).#&S$/",-#%&*0$*& "#$'&*(&*0#&/)./'#)*X&$)'&VDY&*0#&S$,"*&0$'&-#%,"*#'&/)&$)&(*0#-1/%#&,)/)*#)'#'&$)'&,)S(-#%##)&"(%%&(S&./I/"/$)&"/S#<71&="*0(,50&7$)M&(S&*0#&(4#-$*/()%&'/%.,%%#'&

&YALE J. INT’L L. 1, 4 (1984).$

&'$Id.$&($Id.$&)$Id., at 19.$'*$W. Michael Reisman, The Lessons of Qana, 22 YALE J. INT’L L. 381 (1997).$'!$A long list of cases meets the preliminary criteria. This includes, among others, the

US special forces’ helicopters attack in the Uruzgan province of Afghanistan (February 21, 2010); the Turkish Roboski Massacare in Uludere (December 28, 2011); the Dutch airstrike against alleged Islamic State bomb factory in Northern Iraq (June 2, 2015); the U.S. Attack on a Médecins Sans Frontières hospital in Kunduz province of Afghanistan (October 3, 2015); the Saudi Arabian-led coalition Strike on a funeral hall in Sana’a, Yemen (October 8, 2016); and the Nigerian bombing of Rann Refugee Camp (January 17, 2017). Within the limits of this paper, I could not address these cases as well. Nonetheless, the intelligence faults depicted in each of the above cases mirrors many of the faults described in the five selected cases and in that regard do not bear repeating.$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& >D&

/)&*0#&.$%#&%*,'/#%&1#-#&.()',.*#'&/)&$&."$)'#%*/)#&7$))#-X&4(%*3/)./'#)*&/)I#%*/5$*/()%&V:M&:(*0&*0#&*$-5#*/)5&L*$*#X&/)*#-)$*/()$"&(-5$)/`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

Cases selected range temporally from World War II to 2014 and further span multiple continents and belligerent parties, from the U.S. to Germany, to Israel.72 In analyzing the case studies, and where possible, I relied on primary sources, including press releases, official investigations by governmental agencies, legal judgments and rulings, and reports by intergovernmental international organizations. Other parts of the data come from secondary sources, such as press reports and reports from civil society, which sometimes contain unofficial allegations and speculations. The analysis provided below does not aim to offer a full account of all possible case studies that meet the inclusion criteria, nor does it represent the majority of cases or all major cases. To the contrary, these are just a small sample of a possible larger pool, if not ocean of relevant examples. Nonetheless, this body of cases is robust enough to highlight areas of potential correlation and similarity to begin to sketch basic primary standards of care.

The cases examined in this Section are divided between two categories: Intelligence faults resulting from errors in the collection and processing phases, and intelligence faults resulting from errors in the analysis and dissemination phases. Each set of faults triggers different lessons and best practices that must be baked into our evolving duty of care.&

&B. +9;@8B3B&6D&,;B2&.5EF32B

A< B0#&T$**"#&(S&U()*#&8$%%/)(&VR#:-,$-M&AWX&AJOOY&

The Battle of Cassino is the collective name given to a series of strikes

that took place during the first five months of 1944. In each of its four assaults the Allied Forces attempted to break through the German ‘Gustav Line’ by capturing the town and mountain of Cassino. The most tragic and controversial of the four operations was the air campaign launched on

&'"$The research methodology draws some inspiration from Dan Efrony and Yuval Shany,

A Rule Book on the Shelf? Tallinn Manual 2.0 on Cyber Operations and Subsequent State Practice, 112(4) AM. J. INT’L L. 583 (2018).$

>O& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

February 15, against the Abbey of Monte Cassino. This site of cultural and religious importance was founded in 526 by St. Benedict, making it the most ancient house of the Benedictines. The Abbey numbers St. Thomas Aquinas amongst its early monks.73

Even though both sides had given an undertaking to the Pope to keep out of the monastery and protect it, and despite the fact that Germany had declared a ‘Neutral Zone’ of 300 yards around the Abbey in December 1943, many in the Allied forces refused to believe that the Germans would hold true to their word. In reality however, P#)#-$"&R-/'(&I()&L#)5#-&,)'&Z**#-"/)X&*0#&"(.$"&P#-7$)&8(-4%&8(77$)'#-X&1$%&$&bI#-M&'#I(,*&8$*0("/.&$)'&1(,"'&)#I#-&0$I#&:-#$.0#'&%,.0&$)&,)'#-*$2/)5<c!"&P#)#-$"&P#*-/#&B,2#-&*0#&P8H&(S&*0#&O#$&C)'/$)&̂ /I/%/()X&10(&1$%&*$%2#'&1/*0&*0#&.$4*,-#&(S&*0#&7()$%*#-MX&2)#1&)(*0/)5&(S&I()&L#)5#-&$)'&b-#S,%#'&*(&$..#4*&*0$*&*0#&P#-7$)%&1(,"'&%*/.2& *(& *0#& $5-##7#)*c& 4-#'(7/)$)*"M& given the monastery’s strategic natural defensive position.75

It is this political conception that guided intelligence collection in the days leading up to the operation. The British field Marshal, General Harold Alexander, who himself was unsure as to whether the Germans will or already had occupied the Monastery, was the one who set the tone for the entire operation in a note from his staff which read: “If there is any reasonable probability that the building is being used for military purposes, General Alexander believes that its destruction is warranted.”76 It is not uncommon for Intelligence analysts to suffer from over-confidence biases (whereby new information, is evaluated in a way that reasserts and reaffirms previous conceptions. This is caused in part due to the “anchoring effect” phenomenon, which causes people to utilize previous information when processing new information, even when the previous information should be deemed irrelevant).77

It thus should come as no surprise that following General Alexander’s &

'#$For further reading see Nigel De Lee, Moral Ambiguities in the Bombing of Monte Cassino, 4(2) J. MIL. ETHICS 129 (2005); DONALD F. GOBER, BATTLE ANALYSIS-CASSINO: THE SECOND, THIRD, AND FOURTH BATTLES (1984).$

'$$%&$7.-$:1-F1"$'-5$k001"<,-$;,A&1<($1mJ<',-15$<'01"V$d9.-01$C'&&,-.$3.*<5$-171"$;'71$+11-$.//*J,15$+8$gM1"A'-h$'"0,<<1"8$&J.001"###$&.$/.-&J,/*.*&$'$<'-5A'"@$3.*<5$+1$c*,01$ *-&*,0'+<1$ '&$ '-$ .+&1"7'0,.-$ J.&0f$ O>"+23&13*&:#*,#"&(*2&k$$#"'+*=&L#+$-#"&

>#!"&L3"&D30#&RSR&OPT_\UU##'%$ !""# K(0#"$&C'!"7#=&b+$-&%'#8& !$&b!"V&>"3)& $-#& 4""!5!226& $3& $-#&23&

PPT[PYPT[Q&P\QYP\`&ORSSSUe$I.;-$kX'"5=$3&&/&#G"-#./#D/+9)12#/F#8/1."#K'%%)1/=$N-#&

M(!"2+!*$O%J"#$\=$RSSSU=$;00J&VWW333#0;1F*'"5,'-#/.AW3."<5WRSSSW'J"WS[WG.;-1X'"5J$$'&$!""#)11=$%=?&'#-.01$`\=$'0$P\P##''$!""#k53'"5$G#$I.8/1$a$M'"8$C#$?,55<1=$*1A4/&)12#'1-#*-^=%.+"1.#)1#O&/9'9)()%.)A#

$1F"&"1A"# )1# *=-).)12=$ PT$ I#& %//$#& K#.#$ PRS$ OPTZPUe$ %--1$ 7'-$ %'@1-=$ D"4'L)/&'(#$1."&1'.)/1'(#G'I#'1-#3A/1/+)A%=#QQ$D!"1#&4*$n'&)#&I#$[RP=$[RTY[\S#ORSP[U##

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& >W&

general directive, and given the growing belief amongst top Allied military officials that Germans were actually populating the monastery (and more importantly that it is precisely for their use of the Monastery as an artillery observation point that the allied forces were unable to breach the ‘Gustav Line’ thus far),78 intelligence information was soon found to support an operation. One of the main justifications for launching the attack came in the form of a single intercepted communication which was wrongly interpreted, as further summarized by General Clarke of the US 5th Army:

“[A] radio intercept of the German command net reported that a parachute commander had been heard to ask, ‘Ist Abt Kloster?’ and was answerd, ‘Ja in Kloster mit Mönchen.’ The Intelligence Officer who received the intercept only recoded the answer ‘Yes’ from the enemy. The translation then produced was ‘Is the HQ in the Abby?’ the word ‘Abt’ being taken as an abbreviation for ‘Abteil’ (a battalion or unit) rather than Abbot [the head of the monastery]. It was only when Colonel Hunt questioned the translation and the whole intercept that it transpired that the correct reply to the question was in fact Ja in Kloster mit Mönchen,’ i.e. ‘Yes. The Abbot is with the monks in the Monastery. Tragically, this discovery was made too late and the bombers were already approaching... Of course a high percentage did fall on the Monastery itself and reduced much of it to rubble, thereby providing the enemy with a first-class defensive position, of which they took full advantage. Rather than saving lives, that bombing would lead to savage losses and the deaths of many refugees who had been sheltering in the monastery.”79

In early February, days before the decision to launch the attack, an American Air Force general, Ira C. Eaker, flew over the Abby at 200 feet to reconnoiter it. He claimed that he “thought he saw” German military personnel inside the enclosure (“German uniforms hanging on a clothesline in the abbey courtyard; [and] machine gun placements 50 yards from the abbey walls”).80 Major General Geoffrey Keyes of U.S. II Corps flew over the Monastery several times before and after Eaker’s flight. He claimed he saw no evidence of German presence in the Abby, and rejected Eaker’s

&'($!""#)11=$%=?&'#-.01$`\=$'0$P\P##')$!""#C<'"@1=$%=?&'#-.01$`Q=$'0$P\`##(*$!""#!!1+2&K+/-!"2.3*&!*2&!!1+2&D!0,332=&93*$#&C!..+*3V&N-#&:$3"6&34&

$-#&93.$&C3*$"31#".+!'&?!$$'#&34&b3"'2&b!"&44$PZQ$ORSSRUe$9!$$-#5&2!"7#"=&93*$#&C!..+*3V&N-#&D!"2#.$Y>3(,-$&?!$$'#&34&b3"'2&b!"&44$P`\$ORSS\Ue$!""#'(%/#)11=$%=?&'#-.01$`\=$'0$P\R##

>Q& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

account noting that: “They’ve been looking so long, they’re seeing things”.81 But it was not just military officials who pressured this intelligence.

“[A]llied leaders exploited the belief that Monte Cassino was occupied to prepare opinion for aerial attacks on it. On February 11 - four days before the bombing - the Daily Mail ran an army-inspired lead story, Nazis Turn Cassino Monastery into Fort.”82 On February 14 allied guns fired leaflets over the area warning that, in view of the German occupation, "with very heavy hearts we are going to have to turn our weapons on the abbey".83

In total 250 civilians, men, women and children, seeking refuge in the Abbey were killed during the February 15 attack. No German troops were present at the site, and none were directly harmed.84 It was only after its destruction that German paratroopers were ordered to occupy the ruins of the Monastery and turn it into a fortress and observation post, which in turn caused greater disadvantage to the Allied Forces and lingered the Cassino battles. Luigi Maglione, the Pope’s Cardinal Secretary of State summarized perfectly the events that unfolded at Monte Cassino: “a colossal blunder... a piece of a gross stupidity.”85 In light of the growing international criticism General Alexander remained unfazed: “he would simply quote the well-understood principle that no commander could hesitate between destroying a building, no matter how famous or even holy, if it was occupied by the enemy and the lives of his soldiers were thereby put at risk.”86 The land of intelligence blunders is a land of high-handed ostriches.

This case brings to the forefront both mistakes relating to politicized intelligence and groupthink mentality, as well as the potential consequences of reliance on a single data entry (in this case the faulty interpretation of the communication) as the basis for an operation.

2. Operation “Urgent Fury” and the Bombing of the Richmond Hill Mental

Hospital (October 25, 1983) &

The American invasion of Grenada in October 1983, dubbed Operation “Urgent Fury,” was intended to rescue American citizens, restore democratic government, and eliminate Cuban influence on the island.87 As a part of the operation, Navy SEALS were deployed to rescue Grenadian governor-

&(!$!""#K,/;'"5&.-$'-5$D'JF..5=$$-J=$'0$P_T##("$!""#kX'"5=$%=?&'#-.01$`Q##(##$-J#($$$-Je$!""#'(%/=$K,/;'"5&.-$'-5$D'JF..5=$%=?&'#-.01$ZS=$'0$RRQ#$(%$$-J=$'0$RPP##(&$!""#C<'"@1=$%=?&'#-.01$`Q=$'0$P\Q##('$K3*!'2&D#&C3'#=&i0#"!$+3*&6",#*$&>("6V&N-#&2'!**+*,&!*2&k8#/($+3*&34&

I3+*$&i0#"!$+3*.&+*&M"#*!2!#P$OPTT`U#$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& >[&

general Sir Paul Scoon, whom the Americans regarded as the only legitimate governmental authority figure on the island.88 When the SEALS encountered heavy anti-aircraft fire from a Grenadian military installation at Fort Frederick, Vice Admiral Joseph Metcalf – leader of the U.S. Joint Task Force handling operations in Grenada – ordered Navy A-7 jets to bomb all fortified positions in the area.89

Unbeknownst to Metcalf, the fort adjacent to Fort Frederick had been converted from a military base to a civilian mental hospital. 90 Though the site was shown on maps as Fort Matthew, locals knew it as the General Mental Home, or “Crazy House.”91 During the strikes ordered by Metcalf on October 25th, the three-story medical building was hit by a 500-pound bomb, resulting in the death of 21 patients.92 Other, uninjured patients were set loose to wander the streets of the capital.93

Due to restrictions placed on media access to the island, nearly a week passed before news of the hospital bombing broke.94 When it did, reporters questioned why the White House had continued to tout the operation’s lack of civilian casualties during daily briefings after the bombing of the hospital.95 In response, White House spokesman Larry Speakes stated that he did not have adequate information, suggesting that perhaps the Grenadians’ custom of “burying their dead early” had contributed to the U.S.’s slow realization of what had taken place.96 Other White House officials privately acknowledged that the U.S. government had created a credibility problem by limiting media coverage of the invasion.97

Speakes’ statement––that of having inadequate information––aptly characterizes the circumstances surrounding the bombing of the Richmond

&(($?#$!"*AA.-5$%8"1&$I"#=$]J!J#K/1A"-"%#D/+9)12#>/%?).'(#)1#5&"1'-'B#0)(()12#'.#

G"'%.# UQB#L#H#&N+)#.& OL.7#$ P=$ PTZ\U=$ ;00J&VWW333#-80,A1&#/.AWPTZ\WPPWSPW3."<5W*&Y/.-/151&Y+.A+,-FY;.&J,0'<Y,-YF"1-'5'Y@,<<,-FY'0Y<1'&0YPR#;0A<##

()$ 2-+'+0& B(7+#'.7+=& N-#& 6#:#& 4*1!.+3*& 34& M"#*!2!V& )#,!/6& 34& !& >'!5#2&

E+/$3"6&TZYTT&ORSPTU# )*$P$9+/-!#'&9(''+,!*=&M"3(*2&b!"4!"#V&%*&4*$#"*!$+3*!'&k*/6/'30#2+!&

\[P&O:0'-<18$:'-5<1"$15#=$RSSRU#$)!$B*@,1<&@,=&%=?&'#-.01$ZT=$'0$TZYTT#$)"$ $-J&:.A1$"1J."0&$ ,&&*15$ ,-$ 0;1$'(01"A'0;$.($ 0;1$+.A+,-F$"1J."015$'&$A'-8$'&$[`$

51'0;&e$&11$?#$!"*AA.-5$%8"1&$I"#=$]J!J#K/1A"-"%#D/+9)12#>/%?).'(#)1#5&"1'-'B#0)(()12#'.#G"'%.#UQB#L#H#&N+)#.&OL.7#$P=$PTZ\U=$;00J&VWW333#-80,A1&#/.AWPTZ\WPPWSPW3."<5W*&Y/.-/151&Y+.A+,-FY;.&J,0'<Y,-YF"1-'5'Y@,<<,-FY'0Y<1'&0YPR#;0A<#$D.3171"=$1&0,A'01&$,-$0;1$;,F;$011-&$'-5$<.3$031-0,1&$'"1$A.&0$3,51&J"1'5##

)#$B*@,1<&@,=&%=?&'#-.01$ZT=$'0$TZYTT#$)$$$-J$)%$%8"1&=$%=?&'#-.01$ZZ#$)&$$-J#)'$$-J#

>d& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

Hill Mental Hospital. An excerpt from a New York Times article published shortly after the bombing revealed that:

“A map posted in the Public Affairs Office in the Pentagon indicated that the fort complex lies along a ridge of several hundred yards, but none of the buildings are marked as a hospital. The first United States troops ashore in Grenada, it has been reported, did not carry military maps but instead were issued photocopies of tourist maps, apparently the only maps available at the time. It is not known how much detail those maps contained, or whether they showed the fort complex.”98

Indeed after Grenada “moved into the orbit of the Soviet Union and Cuba

in 1979,” the United States had limited ability to collect intelligence and as a result suffered from significant lack of necessary technical intelligence concerning Grenada.99 During an appearance on “Meet the Press” on 6 November l983, “General Vessey candidly summarized URGENT FURY in these words: ‘We planned the operation in a very short period of time--in about 48 hours. We planned it with insufficient intelligence for the type of operation we wanted to conduct. As a result we probably used more force than we needed to do the job, but the operation went reasonably well. ...Things did go wrong, but generally the operation was a success.”100

Indeed, White House and military officials were quick to assert that they did not know the fort contained a civilian hospital.101 Later, they also noted that the building was not marked as a hospital and that gunfire had been directed towards American troops from within the fort.102

On November 5th, nearly two weeks after the bombing, judge advocates and two members of the U.S. Army Criminal Investigation Command (CID) visited the hospital to investigate whether military personnel had violated the Laws of War.103 At the time, the only sign of a violation was an ambiguous “news media report,” made all the more complicated by the lack of a declared state of war between the United States and Grenada or Cuba.104 This investigation revealed that:

“…unlike the roofs of other hospitals in Grenada, there was no red cross marking on the hospital roof. In addition, the walls of the

&)($$-J#))$C.<1=$%=?&'#-.01$Z`=$'0$PQ#$!**$$-#=$'0$_[#$!*!$%8"1&=$%=?&'#-.01$ZZ#$!*"$B*@,1<&@,=&%=?&'#-.01$ZT=$'0$TZYTT#$!*#$>"#2#"+/&)#&?3"/-=&I(2,#&%213/!$#.&+*&C3)9!$&`P&ORSSPU# !*$$$-J$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& >J&

building were marked with the symbols of the People’s Revolutionary Army (large red dots on a white background), and a common wall joined the mental hospital to the revolutionary army headquarters building at Fort Frederick. They also noted that the People’s Revolutionary Army had positioned two antiaircraft batteries approximately fifty meters from the mental hospital, near the nurses’ quarters, and that on 25 October, the day of the bombing, weapons were fired from inside the hospital and from the antiaircraft positions adjacent to its walls.”105

Based on these findings, the investigators concluded that American forces

had not violated IHL. A year after the bombing, the U.S. had yet to publicly apologize or commence restoration of the hospital.106 Ultimately, as Franck summarizes, the U.S. did express regret and offer some compensation “to underscore its assertion that what had occurred was entirely accidental and that no principle was being advanced to defend what had happened.”107 The U.S. military claimed to be conducting Operation Urgent Force with “limited force” and “surgical care.”108 Yet, as this paper argues, a surgeon, or commander, may only be as precise as the tools she relies on.

&D< B0#&=**$.2&()&*0#&80/)#%#&Z7:$%%M&VU$M&[X&AJJJY&

Operation “Belgrade Warehouse 1” was aimed against a Yugoslav

Government agency (“Yugoslav Federal Directorate for Supply and Procurement”, or Yugoimport FDSP) suspected by the CIA of being involved in arms proliferation and procurement. On the Night of May 7, 1999 NATO B-2 stealth bombers fired several missiles hitting the Chinese embassy in Belgrade. Three Chinese citizens were killed and fifteen others were injured. Severe damage was caused to the embassy building.109 The Chinese embassy was mistakenly identified as the Yugoimport FDSP building. The Chinese immediately condemned the strike, calling it a “barbaric attack and a gross

&!*%$$-J$!*&$k<,$C#$91&&,-F1"=$)1001"$0.$0;1$k5,0."=$8"1.'(#>/%?).'(#C"+')1%#'#K'%='(.,#/F#.4"#

5&"1'-'# $1L'%)/1B# L#H#& N+)#.$ OL.7#$ RR=$ PTZ[U=$;00J&VWW333#-80,A1&#/.AWPTZ[WPPWRRW.J,-,.-W<YA1-0'<Y;.&J,0'<Y"1A',-&Y'Y/'&*'<08Y.(Y0;1YF"1-'5'Y,-7'&,.-YPSS[_Q#;0A<#

!*'$k&N-3)!.&9#&>"!*/7=&N-#&235#"&34&)#,+$+)!/6&%)3*,&L!$+3*.&PQP&OPTTSU# !*($?#$!"*AA.-5$%8"1&$I"#=$]J!J#K/1A"-"%#D/+9)12#>/%?).'(#)1#5&"1'-'B#0)(()12#'.#

G"'%.# UQB#L#H#&N+)#.& OL.7#$ P=$ PTZ\U=$ ;00J&VWW333#-80,A1&#/.AWPTZ\WPPWSPW3."<5W*&Y/.-/151&Y+.A+,-FY;.&J,0'<Y,-YF"1-'5'Y@,<<,-FY'0Y<1'&0YPR#;0A<##

!*)$!""#4CNH$kmJ1"0$C.AA,0011$K1J."0=$%=?&'#-.01$Q`=$'0$q$ZS##

Da& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

violation of Chinese sovereignty,” and of “the UN Charter.”110 Following a month-long investigation, Undersecretary of State, Thomas

Pickering, provided oral explanation to the Chinese Government on June 17, 1999. Parts of his explanation are worth quoting in full:

“The bombing resulted from three basic failures. First the technique used to locate the intended target- the headquarters of the [FDSP] was severely flawed. Second, none of the military or intelligence databases used to verify information contained the correct location of the Chinese Embassy. Third, nowhere in the target review process was either of the first two mistakes detected. No one who might have known that the targeted building was not the FDSP headquarters- but was in fact the Chinese Embassy- was ever consulted... In an effort to locate the FDSP building [...] an intelligence officer in Washington used land navigation techniques taught by the U.S. military to locate distant or inaccessible points and objects. These techniques- which involve the comparison of addresses from one street to another- can be used for general geographic location, but are totally inappropriate for precision targeting […] Using this process, the individual mistakenly determined that the building which we now know to be the Chinese Embassy was the FDSP headquarters... Once the wrong target was selected the system of checks that NATO and U.S. command forces had in place to catch target errors did not reveal the mistake. The database reviewers conducted by the European Command were limited to validating the target data sheet coordinates with the information put into the database by [U.S. National Intelligence Mapping Agency] analysts. Such a circular process could not uncover the original error and exposes our susceptibility to a single point of database failure... The bombing of the embassy in Belgrade was a tragic accident occurring during a time of ongoing hostilities in Yugoslavia. While the action was completely unintended, the United States and NATO nevertheless recognize that it was the result of a set of errors which led to the embassy being mistakenly targeted... In view of these circumstances, and recognizing the special status of the diplomatic personnel who were affected, the United States wishes to offer immediate ex gratia payments to those

&!!*$ !""# 3=&/?"# Y*<@# 4).%# K4)1"%"# 3+9'%%,=$ ??C& L#5.$ O9'8$ Z=$ PTTTU=$

;00JVWW-13&#++/#/.#*@WRW;,W1*".J1W\\Z[R[#&0A#$#

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& DA&

individuals who were injured in the bombing and to the families of those killed, based on current experience internationally for the scale of such payments.”111

In even further oversight, CIA Director George Tenet after reviewing the

incident, concluded that “there were people at the CIA and at the Department of Defense who had an intimate understanding of the Belgrade environment, but they were not consulted in this process.”112An Expert Committee was convened at the request of the Prosecutor of the International Criminal Tribunal of the Former Yugoslavia, to examine NATO’s aerial operation. Their report concluded that the U.S. aircrew involved in the attack and the senior leadership who authorized it should not be assigned any criminal responsibility as they were “provided with wrong information”.113 The report further noted:

“NATO, and subsequently various organs of the US Government, including the CIA, issued a formal apology, accepted full responsibility for the incident and asserted that the intended target, the Federal Directorate for Supply and Procurement, would have been a legitimate military objective. The USA has formally apologized to the Chinese Government and agreed to pay $28 million in compensation to the Chinese Government and $4.5 million to the families of those killed or injured. The CIA has also dismissed one intelligence officer and reprimanded six senior managers. The US Government also claims to have taken corrective actions in order to assign individual responsibility and to prevent mistakes such as this from occurring in the future.”114

The U.S. admission of its omissions and its disclosure of its faulty

procedures, despite being brought about through public pressure, is nonetheless admirable. Again, we see mistakes in the processing and decoding phase; that is in the translation of raw data into easily accessible materials for both intelligence analysts and later military commanders and policy makers. What is worrisome, from an accountability perspective, is the Expert Committee’s attitude towards ex gratia (i.e. by favor and not by law) payments, internal admonishments, and diplomatic acts of elastic

&!!!$ !""# :#!*& !#& 9("0-6=& 6*+$#2& :$!$#.& 2"!/$+/#& +*& 4*$#"*!$+3*!'& )!5& o&

E3'()#&PV&PTTTYRSSP=&PSS$O7.<#$4=$RSSRU##!!"$ b+''+!)&9#&%"7+*=&C+1+'+!*&!#!$-.& +*& $-#&L%Ni&%+"&C!)0!+,*V&D()!*&

K+,-$.&b!$/-&K#03"$$QP$ORSSRU#$!!#$!""#4CNH$kmJ1"0$C.AA,0011$K1J."0=$%=?&'#-.01$Q`=$'0$q$ZQ##!!$#$-J=$'0$q$Z[##

D>& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

Apologética. It would seem that these actions, for the Committee are sufficient to exonerate a State and its officials and abrogate any further responsibility. It is perhaps true that the U.S. aircrew and senior leadership are not to blame for they were provided with “wrong information,” but what about the people who provided that information? What liability if any falls on the doorstep of the CIA for its faults and omissions? O< P#-7$)& =**$.2& ()& B$)2#-& B-,.2%& (,*%/'#& *0#& \/""$5#& (S& E$]/& L$20/&

^#':M&/)&_,)',`&!-(I/).#&(S&=S50$)/%*$)&VL#4*#7:#-&OX&>aaJY&&h)'#-&h<e<&L#.,-/*M&8(,)./"&K#%(",*/()&ADdQX&$)&C)*#-)$*/()$"&L#.,-/*M&

=%%/%*$).#& R(-.#& VCL=RY& 1$%& #%*$:"/%0#'& *(& $%%/%*& *0#& =S50$)& C)*#-/7&=,*0(-/*M&/)&*0#&7$/)*#)$).#&(S&%#.,-/*M&/)&_$:,"&$)'&/*%&%,--(,)'/)5&$-#$%<&B0#&P#-7$)&T,)'#%1#0-&4$-*/./4$*#'&/)&*0#&CL=R&7/%%/()<&8("()#"&P#(-5#&_"#/)&.(77$)'#'&*0#&P#-7$)&$-7#'&S(-.#%6&S/#"'&.$74&/)&_,)',`<115&&

F#$'/)5&,4&*(&*0#&#I#)*%&(S&L#4*#7:#-&O#$X&*0#-#&1#-#&/).-#$%/)5&),7:#-&(S&/)%,-5#)*%&0/]$.2/)5&I#0/."#%<&B0#&4-/7$-M&*$-5#*&1$%&4("/.#&$)'&7/"/*$-M&I#0/."#%X&:,*&$"%(&*$)2#-&I#0/."#%<&B0#&4-/7$-M&-/%2&1$%&*0$*&*0#%#&.$4*,-#'&4("/.#&I#0/."#%&1(,"'&:#&,%#'&*(&$**$.2&4#-%())#"&$)'&S$./"/*/#%&(S&*0#&=S50$)&%#.,-/*M& S(-.#%&$)'& CL=R&:M& -#"M/)5&()& %*("#)&,)/S(-7%<& C)'##'X&$&1##2&:#S(-#&*0#&/)./'#)*&$)&$**$.2&,%/)5&$&*$)2#-&(..,--#'&/)&*0#&_$)'$0$-&-#5/()&1/*0&DJ&4#(4"#&'#$'&$)'&QO&/)],-#'<116&

C)& *0#& $S*#-)(()& (S& L#4*#7:#-& O#$& *1(& S,#"& *$)2#-%& :#"()5/)5& *(& *0#&P#-7$)&!-(I/)./$"&K#.()%*-,.*/()&B#$7&1#-#&*$2#)&:M&B$"/:$)&S(-.#%&()&*0#&_,)',`3_$:,"&0/501$M<&b=%& *0#& *$)2#-%&1#-#&.-(%%/)5&_,)',`&-/I#-&:#*1##)&="/&=:$'&$)'&80$-&̂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aXaaa&"/*#-%X&

&!!%#6#L#$:1/#$K1&#$P\Z_=$6#L#$5./#$:WKk:WP\Z_=$q$\Y[$ORSSPU$OdC'<<&$*J.-$91A+1"$

:0'01&$0.$/.-0",+*01$J1"&.--1<=$1c*,JA1-0$'-5$.0;1"$"1&.*"/1&$0.$0;1$4-01"-'0,.-'<$:1/*",08$%&&,&0'-/1$>."/1=$'-5$,-7,01&$0;.&1$91A+1"$:0'01&$0.$,-(."A$0;1$<1'51"&;,J$.($0;1$>."/1$'-5$0;1$:1/"10'"8YM1-1"'<e$%*0;.",X1&$0;1$91A+1"$:0'01&$J'"0,/,J'0,-F$,-$0;1$4-01"-'0,.-'<$:1/*",08$%&&,&0'-/1$>."/1$0.$0'@1$'<<$-1/1&&'"8$A1'&*"1&$0.$(*<(,<$,0&$A'-5'01#fU#$

!!&# :*&'--1$ B'&0-1"=$ D"%A4(=%%"+?F"4(=12# =1-# D"&)A4.=$ P`W`[SS$ !#($./-#"&

?(*2#.$!,=&[R$ORSPPU#$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& DD&

$..(-'/)5&*(&#M#1/*)#%%#%&$)'&I/""$5#-%<c117&C*&/%&$*&*0/%&4(/)*&

*0$*& *0#& B$"/:$)&$""(1#'& *0#& *-$.*(-&'-/I#-%& S-(7&=7#-20/#"& *(&/)S(-7& *0#&-#%/'#)*%& *0$*& *0#M&7$M& %/40()& (SS&%(7#& S,#"<& b=%&1(-'&%4-#$'&/)&*0#&I/""$5#& $)'&#"%#10#-#& $:(,*&*0#& S-##& S,#"X&:#*1##)& Aa<aa& *(&AA<aa& 47& *0#&*$)2#-%& 1#-#&%,--(,)'#'&:M&(I#-&Aaa&7#)&$)'&:(M%&#74*M/)5& *0#& *$)2#-%&(S& *0#& S,#"<&^,-/)5&*0/%&*/7#X&$-7#'&B$"#:$)&S/50*#-%&1#-#&$"%(&4-#%#)*&$*&*0#&%.#)#<c118&&

=-(,)'&*0#&%$7#&*/7#&b*0#&P#-7$)&.$74&5(*&/)S(-7$*/()&S-(7&=S50$)&/)*#""/5#).#&%(,-.#%&*0$*&*0#&*$)2#-%&0$'&:##)&%*("#)&$)'&1#-#&)(1&%*,.2<&=&e=BH& %,-I#/""$).#& $/-4"$)#& '/%.(I#-#'& *0#& "(.$*/()& (S& *0#& *$)2#-%& $)'&:#5$)&*(&%#)'&I/'#(&/7$5#%&*(&*0#&P#-7$)&.$74<&=*&*0#&.$74X&S(,-&P#-7$)&$-7M&7#7:#-%&1$*.0#'&*0#&I/'#(%&$)'&-#.#/I#'&S,-*0#-&/)S(-7$*/()<&B0#&4-/7$-M&%(,-.#&(S&*0/%&/)S(-7$*/()&1$%&()"M&()#&=S50$)&/)*#""/5#).#&(SS/.#-&10(& 1$%& )(*& 4-#%#)*& $*& *0#& %/*#<& E#& -#4(-*#'& I/$& *0-##& (*0#-& 4#-%()%X&/).",'/)5&()#&*-$)%"$*(-X& *0$*&()"M&/)%,-5#)*%&1#-#&4-#%#)*&$*& *0#&%/*#<c119&B0/%&5$7#&(S&b:-(2#)&*#"#40()#c&:#*1##)&*0#&/)S(-7$)*&$)'&*0#&P#-7$)&(SS/./$"%&1(,"'&4(%#&()#&*0#&4-/7$-M&-#$%()%&S(-&*0#&/)./'#)*<&&

_"#/)&1$%&/)&.()%*$)*&.()*$.*&1/*0&*1(&=7#-/.$)&RAW&]#*%X&:,*&e=BH6%&-,"#%& (S& #)5$5#7#)*& $""(1& S(-& :(7:/)5& ()"M& 10#)& CL=R& S(-.#%& $-#&#)'$)5#-#'<&B0#&4/"(*%&-#4#$*#'"M&$%2#'&_"#/)&/S&*0#-#&1$%&$)M&*0-#$*&4(%#'&

&!!'# %A-1&08$ 4-01"-'0,.-'<=$ *F24'1)%.'1:# D'A[2&/=1-# ./# .4"# 0=1-=6# ')&%.&)["# /F# _#

!"?."+9"&# QTT`=$ PYR# Oi/0#$ \S=$ RSSTU=$;00J&VWW333#'A-1&08#."FW5.3-<.'5W!./*A1-0&W[[SSSW'&'PPSPQRSST1-#J5(#$

!!(#$-J$!!)# k*".J1'-$ C1-01"$ (."$ C.-&0,0*0,.-'<$ '-5$ D*A'-$ K,F;0&$ OkCCDKU=$ 5"&+'1# *)&#

!.&)["#1"'&#0=1-=6aa*#Z"'&#*F."&:#3L'(='.)/1#/F#̂ =-)A)'(#&"'A.)/1%#'1-#F=&.4"&#)1F/&+'.)/1=$\$ O%*F#$ \S=$ RSPSU$ ;00JVWW333#'5;YF1-171#/;WK6)%CW-13&WkCCDKYB*-5*XY%YH1'"Y%(01"#J5(#$

&C"",%*-$*/()&A9&80$/)&(S&8(77,)/.$*/()%&&

/)&*0#&P#-7$)&=**$.2&

DO& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

*(&P#-7$)&S(-.#%<%&'&E#&/)%/%*#'&*0$*&*0#-#&1$%&%,.0&*0-#$*&S-(7&*0#&I#0/."#%&*(&*0#&P#-7$)&.$74&#I#)&*0(,50&*0#&*-,.2%&1#-#&%*,.2&()&$&%$)':$)2&/)&*0#&-/I#-<&B0#M&S,-*0#-&$%2#'&10#*0#-&*0#M&%0(,"'&S"M&$*&"(1&$"*/*,'#&$%&$&%0(1&(S& S(-.#& $)'&1$-)/)5<& _"#/)&'#."/)#'X& %,55#%*/)5&*0$*& '(/)5& %(& 1(,"'&$"#-*& *0#& B$"/:$)& (S&*0#& CL=R& S(-.#%6&4-#%#).#<%&%&R(""(1/)5& *0/%&#f.0$)5#&*0#&RAW&]#*%&'-(44#'& *1(& PT=3AW& :(7:%& ()& *0#&.-(1'<& bB0#& :(7:%&0/*& 7(-#& *0$)& Aaa&4#(4"#X&7(%*&(S&*0#7&./I/"/$)%<& =7()5&*0#7& 1#-#& $&.()%/'#-$:"#& ),7:#-&(S&.0/"'-#)<c%&&&

_"#/)&1(,"'&"$*#-&5(& ()& *0#& -#.(-'& *(&%*$*#& *0$*& :$%#'& ()& 0/%& /)*#-$.*/()& 1/*0& *0#& %/)5"#& EhUCeB& %(,-.#X& bC&$%%,7#'&*0$*&$""&*0#&4#(4"#&'(1)&*0#-#&:#"()5&*(&*0#&/)%,-5#)*%<&C&.$))(*&%$M&10#*0#-& *0#M&1#-#&0("'/)5& *0#/-&5,)%& /)& *0#/-&0$)'%&(-&4,**/)5& *0#7&'(1)<& C& 0$'& )(& '(,:*& $:(,*& *0/%& -#4-#%#)*$*/()<& C& $"%(& +,#%*/()#'& *0/%&4-#%#)*$*/()&%#I#)&*/7#%&$)'&/*&0$%&:##)&.()S/-7#'&*(&7#&7$)M&*/7#%<c%&(&&

B0#&/)./'#)*&*-/55#-#'&$&),7:#-&(S&.(,-*&4-(.##'/)5%&$)'&4$-"/$7#)*$-M&/)I#%*/5$*/()%&/)&P#-7$)M<&B0(%#&"#5$"&4-(.#%%#%&'#7()%*-$*#'&S,-*0#-&*0$*&/)&*0#&"#$',4&*(&*0#&$**$.2&*0#&%$)':$)2&bwas continuously monitored by the two fighter planes for at least 20 minutes. The infrared recordings made by the on-board cameras were transferred to the air traffic control center in real time.”124 The commanders relied on those infrared images to reach their

&!"*$$-#$!"!$%A-1&08$4-01"-'0,.-'<=$%=?&'#-.01$PP`=$'0$RY\#$!""$kCCDK$K1J."0=$%=?&'#-.01$PPT=$'0$\#$!"#$B'&0-1"=$%=?&'#-.01$PP_#$!"$$?.--$j,7,<@'AA1"=$D"%."4"1#L/1#!A4'-"1%"&%'.6'1%?&bA4"1#'=%#-"+#L/1#")1"+#

-"=.%A4"1# D=1-"%I"4&R*12"4c&)2"1# '+# T_JT`JQTT`# 9"F/4("1"1# G=F.'1R2&)FF# '=F# 6I")#<'1[('%.6b2"#)1#-"&#Yd4"#L/1#0=1-=6#)1#*F24'1)%.'1$OL.7#$PR=$RSP\U#$

&C"",%*-$*/()&>9&F/I#&R((*$5#&S-(7&*0#&^-()#%&/)&

P#-7$)&=**$.2&&&

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& DW&

determinations of the evolving situation on the ground. Together with the pilot-in-command officer (forward air controller) they attempted to evaluate what they were seeing on the ground.125 The infrared imagery did not capture any useful information about the targets (namely height, in order to determine that certain of the targets were indeed children). Individuals were only identified as “points” or “dots.

Germany ultimately went ahead and paid compensation to the families at the amount of 5,000 USD per victims’ family as a lump sum ex gracia payment. The German Constitutional Court declined to hear cases brought against Klein and the other military members. Nonetheless, questions as to the unavoidability of the intelligence failures in this incident—the inadequate way that information was transferred along the chain, the overconfidence of the intelligence assessors, the inability to vet the information received—continue to linger.&&W< H4#-$*/()&b!-(*#.*/I#&Z'5#c&$)'&*0#&P$`$&T#$.0&=**$.2&V@,"M&AQX&>aAOY&

Operation “Protective Edge” took place from July 7 to August 26, 2014. It was yet another round of intense hostilities between Israel and Hamas and other various armed groups. The first phase of the operation involved an extended aerial campaign launched by Israel with the purpose of degrading the military capacity of Hamas and the other organizations in the Gaza Strip, and in particular halting or significantly reducing the organizations’ rocket launching.126 One of the most controversial aerial strikes conducted during Israel’s campaign resulted in the death of four children aged 9-11 in one of Gaza’s beaches. The attack, which occurred on July 16th, 2014, immediately sparked international controversy, in part due to the particular beach’s close proximity to a hotel housing international journalists.

The Israeli Military Advocate General’s (MAG) office led an internal investigation of the incident. The investigation concluded that the children were killed from the Israeli strike. Nonetheless, the MAG ordered on June 11, 2015, that “the investigation file be closed without any further legal proceedings – criminal or disciplinary – to be taken against those involved in the incident”. Lt. Col. Peter Lerner, the Israeli Defense Forces Spokesman published an official statement summarizing the findings of that investigation which I provide in verbatim:

“The incident took place in an area that had long been known as a compound belonging to Hamas's Naval Police and Naval Force

&!"%$$-J$$!"&#!""$4&"'1<$RSP[$M'X'$C.-(<,/0$K1J."0=$%=?&'#-.01$RZ=$'0$q$ZR##

DQ& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

(including naval commandos), and which was utilized exclusively by militants. The compound in question spans the length of the breakwater of the Gaza City seashore, closed off by a fence and clearly separated from the beach serving the civilian population. It further found in the course of the investigation (including from the affidavits provided by Palestinian witnesses), that the compound was known to the residents of the Gaza Strip as a compound which was used exclusively by Hamas's Naval Police. The IDF carried out a number of attacks on the compound in the days prior to the incident. In the course of one such attack, which took place on the day prior to the incident (15 July 2014), a container located inside the compound, which was used to store military supplies, was attacked. Shortly before the incident, an intelligence assessment was established which indicated that operatives from Hamas's Naval Forces would gather in the military compound in order to prepare for military activity against the IDF. On 16 July, aerial surveillance identified a number of figures entering the compound at a running pace. These figures entered a shed adjoining the container which had been attacked the day prior. Against the backdrop of the aforementioned intelligence assessment, these were believed to be militants from Hamas's Naval Forces, who had arrived at the compound in order to prepare to execute the aforementioned military activity against the IDF. It should be stressed that the figures were not identified at any point during the incident, as children. In light of the above, it was decided to conduct an aerial attack against the figures which had been identified, after all the necessary authorizations for an attack had been obtained, and after a civilian presence in the area had been ruled out. When one of the identified figures entered into the remains of the container which had been attacked on the day prior to the incident, one missile was fired from the air towards the container and the adjoining shed. As a result of this attack, it appeared that one of the figures identified was hit. Following this attack, the rest of the figures began to run in the direction of the compound's exit. Shortly before their exit from the compound, an additional missile was fired from the air towards them, which hit the figures in question after they had exited the compound. Tragically, in the wake of the incident it became clear that the outcome of the attack was the death of four children, who had entered the military compound for reasons that remain unclear. It further arose from the investigation that, under the circumstances in question, it would not have been possible for the operational entities involved to have identified these figures, via aerial surveillance, as children….

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& D[&

At the time that the decision was made, the attack was not, according to the assessment of the operational entities, expected to result in any collateral damage to civilians or to civilian property…. However, it became clear after the fact that the identification of the figures as militants from Hamas's Naval Forces, was in error. Nonetheless, the tragic outcome of the incident does not affect the legality of the attack ex post facto. Accordingly, the Military Advocate General ordered that the investigation file be closed without any further legal proceedings–– criminal or disciplinary––to be taken against those involved in the incident. Nonetheless, inter alia as a result of this incident, the IDF has been working to improve a number of its operational capabilities, including technological capabilities, in order to minimize the risk of the recurrence of tragic incidents of this kind.”127

Not all agree with this narrative of the attack and the events that preceded

it.128 However, even were we, for the purposes of this analysis, to assume that this is an accurate depiction of the entire incident, we will still encounter a series of intelligence faults in the assessment stage: beginning with an erroneous intelligence assessment which anticipated a meeting of Hamas naval operatives that never took place; followed by a misguided determination that the compound was vacant of civilian presence both generally and at the time the operation was authorized; and closing with a problematic identification of four young children as Hamas militants.

Without being provided more information as to the way intelligence was collected and analyzed, there are reasons to assume that some form of groupthink and over-confidence biases played a role in these faults. However, more dramatically, I believe, this case brings to the forefront two primary root causes for intelligence analysis deficits in the context of aerial targeting. The first concerns an over-reliance on “aerial surveillance” as a primary source for intelligence leading up to and during the operation. The Israeli official noted that: “it would not have been possible for the operational entities involved to have identified these figures, via aerial surveillance, as children“ and that “the attack was carried out while undertaking several precautionary measures, [...including] the deployment of real time visual

&!"'# !""$ )0#$ C.<#$ 2101"$ )1"-1"=$ >'/1+..@$ 2.&0,-F$ OI*-1$ PP=$ RSPQU=$

;00J&VWW333#('/1+..@#/.AW)0#C.<#2101")1"-1"WJ.&0&WT``QQ[RTZTQRPT_VS#$ !""# '(%/#iJ1"'0,.-$2".01/0,71$k5F1V$4-71&0,F'0,.-$.($1m/1J0,.-'<$,-/,51-0&$Y$6J5'01$[$

!"(#>."$ '$ &*AA'"8$ .($ &.A1$ .($ 0;1$ /.-0"'5,/0,-F$ 7,13&$ &11$ 2101"$?1'*A.-0=$ $%&'"(#"P/1"&'."%#).%"(F#/L"&#5'6'#9"'A4#[)(()12#/F#F/=&#A4)(-&"1#('%.#,"'&=$N-#&M(!"2+!*$OI*-1$PP=$ RSPQU$ 'L')('9("# '.$ ;00JVWW333#0;1F*'"5,'-#/.AW3."<5WRSPQWG*-WPPW,&"'1<Y/<1'"&YA,<,0'"8YF'X'Y+1'/;Y/;,<5"1-#$

Dd& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

surveillance”. The official thus seems to think that these two facts vindicate Israel from liability. I beg to differ, if it is the indeed the case that the identity of the children could not be established solely on the basis of aerial surveillance, then until the technology advances “aerial surveillance” simply cannot be the only basis for a verification scheme that meet IHL’s principles.

This ties in to the second issue, concerning “signature strikes”. This American term refers to the targeting of individuals solely on the basis of their “intelligence ‘signatures’ – patterns of behavior that are detected through signals intercepts, human sources, and aerial surveillance”.129 As Spencer Ackerman from The Guardian explains, the problem is that:

“[The] signatures at issue are indicators that intelligence analysts associate with terrorist behavior- in practice, a gathering of men, teenaged to middle-aged, traveling in convoys or carrying weapons. In 2012, an unnamed senior official memorably quipped that the CIA considers “three guys doing jumping jacks” a signature of terrorist training”.130

The Israeli official makes sure to include in his public statement that the

group of “figures” entered the compound “at a running pace”. Runs, jumping jacks, and skips cannot be the basis for a decision to launch an aerial targeting operation. Reliance on such circumstantial evidence derived often predominantly from VISINT, does not meet the “heavy” evidentiary burden of proof that Barak set to establish in the Targeted Killings case. Combined the reliance on a single source (a muted aerial imagery) in an attempt to derive circumstantial “patterns of behavior” almost guarantee the formation of misconceptions and biases that lead to insufficient verification.

CCC<& K(&*,+/0.$/,0)C,#(''/0(,&()/,)N%$

The above case studies all illustrate the pivotal role that intelligence plays

in wartime decision-making for targeting purposes. In all of the scenarios examined failures of verification of information, faults in processing and analysis, and an over-reliance by military commanders and intelligence officers alike on a few questionable sources, seem to have been at the heart of the human errors that caused these tragedies.

Undoubtedly, all military commanders and all intelligence officers are &

!")# !""$ M"1F$ 9,<<1"=$ ;4)."# >/=%"# '??&/L"%# 9&/'-"&# Z"+"1# -&/1"# A'+?')21=$ N;1$b'&;,-F0.-$2.&0$O%J"#$RQ=$RSPRU$'L')('9("#'.$;00J&VWWF..#F<W6:!8P)#$

!#*#!""$:J1-/1"$%/@1"A'-=$$1%)-"#@9'+'e%#W&/1"#O'1/?.)A/1:#*#!"A&".#8'A4)1"#;).4#Y/#*AA/=1.'9)().,B#N;1$M*'"5,'-$O%J"#$RQ=$RSPQU$'L')('9("#'.$333#0;1F*'"5,'-#/.AW*&Y-13&WRSPQW'J"WRQW*&Y5".-1YJ".F"'AY&1/"1/8Y&/"*0,-8Y&,F-'0*"1Y&0",@1&$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& DJ&

prone to such errors, but the cases further exemplify a troubling trend: the increasing role of surveillance technology and emerging means of communication as yet a new source for these errors. The process of translating and transferring information from one intelligence actor to another and to a policy maker, from machine to person, and then from person to person––is where many of the errors discussed have materialized.%(%

The idea that “today’s war pilots do not necessarily see the target, and their bombs strike targets whose coordinates have been pre-programmed,” reflects an automated battlefield that only exacerbates the likelihood of mistake.%(& What more, intelligence communities are run on the basis of a complex network of systems, agencies, and organizations (both domestic and transnational). Such networks only compound problems and open the door for even further mistakes.

If one thing stands out from the discussion in Section II it is that the existing internal supervisory mechanisms of the military commander are unlikely to succeed in achieving optimal deterrence and in mitigating risk. All of the aerial strikes discussed above should and could have been avoided had better practices been employed on the part of the intelligence agencies in question. These practices are ones that military commanders have limited capacity to manage on their own. This is because intelligence professionals form a guild, an association of crafts men and women with their own tools, language, and skills.

Indeed, coding and encoding quantitative and qualitative intelligence threat and target assessments is an art form, and like any art form, it follows an internal codebook: a set of practices, procedures, and acronyms known only to members of the trade. Like any other profession, its members owe a duty of care to the societies they serve, the targets they surveil, and the international community writ large. Developing this duty of care, what I call a “reasonable intelligence analyst” test, is but one of the many tools at the disposal of the diligent regulator to manage the otherwise dangerous behavior of these specialized professionals and their hazardous activity. And just like any introduction of a new duty of care we must apply a basic cost-benefit analysis to determine whether such a duty would be efficient in deterring unnecessarily risky conduct. Our ultimate goal is therefore to ensure the adoption of all available socially beneficial preventative measures.

TM&4-(4(%/)5&$)&#f4$)%/()&()&#f/%*/)5&"/$:/"/*M&S-$7#1(-2%X&*0/%&L#.*/()&/%&/)&'/-#.*&.()I#-%$*/()&1/*0&(*0#-&%.0("$-%&10(&0$I#&"((2#'&$*&*0#&%*$*#&(S&

&131$!""#"J2J#9#C#$k<,&;=$8/&'(#K&=+?("#f/1"%:#K'=.)/1'&,#<'("%# )1#>=+'1RC/9/.#

$1."&'A.)/1=$ ;00J&VWWJ'J1"&#&&"-#/.AW&.<\WJ'J1"&#/(A]'+&0"'/0r,5^R`Q`R\_$ O5,&/*&&,-F$&,0*'0,.-&$3;1"1$A,&/.AA*-,/'0,.-&$+10311-$;*A'-$+1,-F&$'-5$A'/;,-1$&8&01A&$"1&*<015$,-$J"171-0'+<1$'//,51-0&U#$!

132$!""#4CKC$kmJ1"0$9110,-F$K1J."0=$%=?&'#-.01$RR=$'0$QT#!

Oa& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

CEF& -#5/7#%& $)'& #I$",$*#'& *0#& 4(%%/:/"/*M& (S& -#S(-7& *0-(,50& *(-*/(,%&"/$:/"/*M<%((& The following pages complement the existing literature by proposing a new duty of care for intelligence officers and those who employ them, as an important piece in a broader reform of IHL accountability. According to this argument Article 57 of API should be reinterpreted and expanded to incentivize intelligence agencies to better their internal intelligence processing and assessment structures and information sharing practices. If Article 57 is read as requiring certain set of expected behaviors, then a failure to meet that standard would result in a serious violation of IHL. Such a violation could trigger the responsible State’s liability under the ILC Articles on State Responsibility.

The Section moves progresses this argument in the following order, first I discuss the potential value in recognizing a new duty of care. Second, I describe the scope and nature of this new duty. Finally, I propose how this duty could be read into existing IHL frameworks and further embedded in informal IHL law-making moving forward.

+/&&G;H394&512&,;B2&D6<&;&%2=&(E58&6D&,;<2&D6<&$952@@3429A2&&

A< =''-#%%/)5&*0#&=..(,)*$:/"/*M&\$.,,7&S(-&N$-*/7#&C)*#""/5#).#&&In her most recent work Mary Manjikian illustrates a repeated syndrome

in the work of the U.S. executive, one she describes as: “rescuing the state by blaming the intelligence community.”134 She cites to multiple occasions where U.S. presidents, from Reagan, to Clinton, to Bush all claimed that they were blindsided by operations of the intelligence community where they were “unaware of the exact nature” of the activities, claiming that if they knew

&!##$ !""B# "J2J=$ %-5"13$ C;,<51"&$ a$ %--'$ )'A*0=$ G"2'(# 7/=1-'.)/1%# F/&# g8'[)12#

*+"1-%h# ./#K)L)()'1%#>'&+"-#9,#*&+"-#K/1F()A.=$D!"1#&4*$n'&D()!*&K+,-$.&C'+*+/$ORSPRU=$ '0$ RY\=$;00JVWW333#-*;'-.7,/(.*-5'0,.-#."FW*&1"W(,<1WRSPRr/;,<51"&r'-5r<'A*0r.-r<1F'<r(.*-5'0,.-&r(."rA'@,-Fr'A1-5&r0.r/,7,<,'-&r;'"A15r+8r'"A15r/.-(<,/0#J5(e$ H'1<$ K.-1-=$*L/)-#/&#K/+?"1%'."i#G)'9)().,#F/&#$1A)-"1.'(#$1^=&,#./#K)L)()'1%#$1F()A."-#W=&)12#*&+"-#K/1F()A.=$ [R$E!*2#& I#&N"!*.*!$n'&)#$ PZP$ ORSSTUe#!,101"$ ><1/@=$ $1-)L)-='(# '1-# !.'."#C"%?/1%)9)().,#F/&#H)/('.)/1%#/F#.4"#$=%#)1#D"((/:#*1#$+?"&F"A.#D'('1A"=$)1#4*$#"*!$+3*!'&

D()!*+$!"+!*&)!5&>!/+*,&L#5&C-!''#*,#.$ORSS`Ue$ K1+1//'$ C".0..(=$ ;'&# </&.%:#*AA/=1.'9)().,#F/&#*=./1/+/=%#;"'?/1%=$P_[O_U$6*+1#&2!#&)#&K#1#$P\[`e$D',A$%+"';'A=$</&.#G)'9)().,#F/&#D"(()2"&"1.#;&/12%=$\T$i843"2&I#&)#,#&:$(2#$ZSZ$ORSPTUe$kA'-*1<'YC;,'"'$M,<<'"5=$C"?'&'.)/1# F/&# H)/('.)/1# /F# $1."&1'.)/1'(#>=+'1).'&)'1# G'I=$ ZQ& 4CKC&

K#1#$QRT$ORSS\U#$134 MARY MANJIKIAN, GENDER SEXUALITY AND INTELLIGENCE STUDIES: THE SPY IN

THE CLOSET 252 (2020).

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& OA&

they “would have taken steps to see that such events did not occur.”135 In so doing, each President excused himself from blame, laying it instead “on the doorstep of the intelligence community.”136

Here lies the danger: If citing faulty intelligence and faulty intelligence operations is enough to establish the existence of a wartime accident, of the kind that triggers a blanket safe harbor from liability, what stops such excuses from forever being cited? Taking the argument to its extreme would completely empty from meaning the object and purpose of IHL, to limit the effects of war on people and property, including those individuals and groups that are particularly vulnerable.

Introducing a duty of care is thus necessary to enhance the enforcement capacity of IHL by preventing the ability of States from claiming intelligence errors as a get out of jail free card. Such a duty would further help clarify the distinction between what Marko Milanovic called “honest but unreasonable” mistakes in war and those which are “honest but reasonable.”137 It would help clarify that not all intelligence blunders are by default “reasonable” and will help elucidate if, when, and under what circumstances, the behavior of the intelligence community might be so egregious as to trigger a violation of the principle of precautions in attack further resulting in state responsibility for an internationally wrongful act. As Marko notes such a violation would be a “plain, ‘vanilla’ violation of IHL,” not to be confused with a war crime.138

&>< Z)0$)./)5&C)*#-)$*/()$"&F$16%&Zf4-#%%/I#&\$",#&

&=%&^$)&_$0$)&1-(*#&b4,)/%07#)*&/%&)(*&],%*&$&1$M&*(&7$2#&(SS#)'#-%&

%,SS#-i& /*& /%& $& %4#./$"& %(./$"& .()I#)*/()& *0$*& %/5)/S/#%& 7(-$"&.()'#7)$*/()<c139&L/7/"$-"M&$)'&L,)%*#/)&)(*#'&*0$*9&b!("/.M&.0(/.#%&'(&)(*&],%*&:-/)5&$:(,*&.#-*$/)&/77#'/$*#&7$*#-/$"&.()%#+,#).#%i&*0#M&$"%(&1/""&:#&,)'#-%*(('X&$*& */7#%X& *(&:#& /74(-*$)*& S(-&10$*& *0#M&-#S"#.*&$:(,*&I$-/(,%&I$",#&.(77/*7#)*%j$:(,*&10/.0&I$",#%&*$2#&4-/(-/*M&(I#-&(*0#-%&(-&0(1&I$-/(,%&I$",#%&$-#&:#%*&,)'#-%*(('<&T(*0&*0#&7$*#-/$"&.()%#+,#).#%&$)'&*0#&#f4-#%%/I#& .()%#+,#).#%& (S& 4("/.M& .0(/.#%& $-#& $44-(4-/$*#& .().#-)%& S(-&

&135 Id. 136 Id. 137 Marko Milanovic, Mistakes of Fact When Using Lethal Force in International Law:

Part I, EJIL: TALK! (Jan. 14, 2020), https://www.ejiltalk.org/mistakes-of-fact-when-using-lethal-force-in-international-law-part-i/.

138 Id. 139 Dan M. Kahan, What do Alternative Sanctions Mean?, 63 U. CHI. L. REV. 591, 599

(1996).

O>& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

4("/.M7$2#-%<c140&C)& (*0#-& 1(-'%X& "$1& 4"$M%& $)& #f4-#%%/I#& $)'& 4-(]#.*/I#& S,).*/()<&

K#5$-'"#%%&(S&$.*,$"&#)S(-.#7#)*X&$-*/.,"$*/)5&-,"#%&*0$*&1(,"'&'/-#.*"M&:/)'&*0#&(4#-$*/()%&(S&*0#&/)*#""/5#).#&.(77,)/*MX&1(,"'&7$-2&$&%/5)/S/.$)*&%0/S*<&C)&*0#&7(%*&/77#'/$*#&"#I#"X&/*&1/""&#)*$/"&*0$*&7/"/*$-M&/)*#""/5#).#&(SS/.#-%&1/""& :#& *-$/)#'X& 4#-0$4%& S(-& *0#& S/-%*& */7#X& ()& *0#& -,"#%& (S& *0#& P#)#I$&8()I#)*/()%X&10/.0&$44"/.$*/()&*0#M&1#-#&%(&S$-&/)%,"$*#'&S-(7<&&

T,*&#I#)&7(-#&S,)'$7#)*$""MX&*0#&#f4-#%%/I#&S,).*/()&(S&*0#&"$1&.(,"'&4"$M&$&"()53*#-7&-("#&/)&-#$'],%*/)5&(,-&)(-7$*/I#&%*-,.*,-#%&/)&*0#&S$.#&(S&$&.(""#.*/I#&$.*/()&4-(:"#7<&=%&L,)%*#/)&'#%.-/:#%9&

&bC7$5/)#X&S(-&#f$74"#X&*0$*&*0#-#&/%&)(&)(-7&/)&S$I(-&(S&-#S,%/)5&*(&"/**#-X&(-&*0$*&*0#-#&/%&#I#)&$&)(-7&/)&S$I(-&(S&"/**#-/)5<&C)&*0#&S$.#&(S&%,.0&)(-7%X&*0#&%(./$"&7#$)/)5&(S&"/**#-/)5&7$M&:#&/)'#4#)'#).#&$)'&S#$-"#%%)#%%X&$)'&*0#&%(./$"&7#$)/)5&(S&."#$)/)5&,4&(-&S$/"/)5&*(&"/**#-&7$M& :#& S$%*/'/(,%)#%%& (-& #I#)& .(1$-'/.#& (-& )#,-(%/%<& C)& %,.0& $&%/*,$*/()&$&%(./#*M&1(,"'X&,)'#-&/7$5/)$:"#&$%%,74*/()%X&'(&1#""&*(&-#.()%/'#-&$)'&-#.()%*-,.*&/*%&)(-7%<&C*&7$M&:#&$:"#&*(&'(&%(&*0-(,50&I(",)*$-M&#SS(-*%<&C)'##'X&)(-7&#)*-#4-#)#,-%&/)&*0#&4-/I$*#&%40#-#&$**#74*&*(&.0$)5#&)(-7%&:M&/'#)*/SM/)5&*0#/-&:$'&.()%#+,#).#%&$)'&*-M/)5&*(&%0/S*&*0#&:$%#%&(S&%0$7#&$)'&4-/'#k&T,*&%(7#*/7#%&*0#%#&4-/I$*#&#SS(-*%&S$/"<&N0#)&*0/%&/%&%(X&*0#&"$1&7/50*&:#&#)"/%*#'&$%&$&.(--#.*/I#<& C)& S$.*X& *0#& "#$%*& .()*-(I#-%/$"& ,%#& (S& *0#& #f4-#%%/I#&S,).*/()&(S&"$1&(4#-$*#%&/)&*0/%&1$M<&E#-#&*0#&5($"&/%&*(&-#.()%*-,.*&#f/%*/)5&)(-7%&$)'&*(&.0$)5#&*0#&%(./$"&7#$)/)5&(S&$.*/()&*0-(,50&$&"#5$"&#f4-#%%/()&(-&%*$*#7#)*&$:(,*&$44-(4-/$*#&:#0$I/(-<c141&

&=&)#1&/)*#-)$*/()$"&%*$)'$-'&S(-&1$-*/7#&/)*#""/5#).#&(4#-$*/()%&.(,"'&

4"$M&$&S(-.#&7,"*/4"/#-&S(-&()5(/)5&'#:$*#%&$:(,*&*0#&-("#&(S&/)*#""/5#).#&/)&%(./#*M<142& C*& .(,"'& S,-*0#-& %#*& $& S(-,7& S(-& .()I#-%$*/()%X& *0$*& 1#-#&4-#I/(,%"M& ,)/7$5/)$:"#& :#*1##)& /)*#""/5#).#& $)'& )$*/()$"& %#.,-/*M&4-(S#%%/()$"%& ()& *0#& ()#& 0$)'X& $)'& 0,7$)& -/50*%& $)'& 0,7$)/*$-/$)&/)*#-)$*/()$"/%*%& ()& *0#& (*0#-<& L,.0& .()I#-%$*/()%& .(,"'& 0$I#& )(-7&

&140 Richard H. Pildes and Cass R. Sunstein, Reinventing the Regulatory State, 62 U. CHI.

L. REV. 1, 66 (1995). 141 Cass R. Sunstein, On the Expressive Function of the Law, 144 PA L. REV. 2021,

2030-31 (1996). 142 For more on these ongoing debates, see Asaf Lubin, SolarWinds as a Constitutive

Moment: A New Agenda for the International Law of Intelligence, JUST INTELLIGENCE (Dec. 23, 2020), https://www.justsecurity.org/73989/solarwinds-as-a-constitutive-moment-a-new-agenda-for-the-international-law-of-intelligence/.

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& OD&

/)*#-)$"/`/)5& #SS#.*%X&10/.0& *0#7%#"I#%& .(,"'& "#$'& *(&10$*& L,)%*#/)& .$""%&b)(-7& .$%.$'#%X& $%& -#4,*$*/()$"& /).#)*/I#%& %0/S*& :#0$I/(-& /)& )#1&'/-#.*/()%<c143&

&D< ^#I#"(4/)5&U#.0$)/%7&S(-&K#'-#%%&

&B0#& -,"#& *0$*& $& %*$*#& b-#%4()%/:"#& S(-& I/("$*/()%& (S& /)*#-)$*/()$"&

0,7$)/*$-/$)&"$1&/%&-#+,/-#'&*(&7$2#&S,""&-#4$-$*/()&S(-&*0#&"(%%&(-&/)],-Mc&/%&$&)(-7&(S&.,%*(7$-M&/)*#-)$*/()$"&"$1&-#.(5)/`#'&/)&:(*0&/)*#-)$*/()$"&$)'&)()3/)*#-)$*/()$"&$-7#'&.()S"/.*%<144&&

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

&bN0#*0#-&(-&)(*& /*%& $.*/()%&1#-#& /)*#-)$*/()$""M&.-/7/)$"&(-&1#-#&7$-2#'&:M&$&.0$/)&(S&5-/#I(,%&#--(-%X&/))(.#)*&I/.*/7%&$-#&#)*/*"#'&*(&*0#&-#4$/-&(S&*0#/-&/)],-/#%<&B0$*&-#4$/-&%0(,"'&)(*&.(7#&S-(7&*0#&/)*#-)$*/()$"&.(77,)/*MX&:,*&S-(7&*0#&4$-*M&*0$*X&$-5,$:"M&/)&1$M%&.(74$*/:"#&1/*0&*0#&"$1&(S&1$-X&#"#.*#'&*(&-#',.#&/*%&(1)&#f4(%,-#&$)'&.()*$/)&/*%&(1)&/)],-/#%&:M&%0/S*/)5&*0#&'$)5#-&$)'&.()%#+,#)*&/)],-M&()*(&(*0#-%<& C)& /74(%/)5& *0#& .(%*%& (S& *0#& #)5$5#7#)*& ()*(&/))(.#)*&4$-*/#%X&$&:#""/5#-#)*&%0(,"'X&"/2#&$)M()#&/)&1$-&(-&4#$.#X&:#& (:"/5#'& *(& -#4$/-& *0#& /)],-/#%& /*& 0$%& .$,%#'& /)& $& 7#$%,-#&4-(4(-*/()$"&*(&/*%&.()*-/:,*/()&*(&*0#7<c145&&

&e(*&()"M&*0$*&:,*&$&%*$)'$-'&*0$*&0("'%&*0#&L*$*#&*(&$..(,)*&S(-&$:,%#%&

.(77/**#'&:M& /*%& /)*#""/5#).#&.(77,)/*M&.(,"'&4"$M&$)& /74(-*$)*& -("#& /)&#)0$)./)5& $& %#)%#& (S& ],%*/.#<& B0#& )(*/()& *0$*& *0#& *$-5#*%& (S& .(I#-*&/)*#""/5#).#&(4#-$*/()%&$-#&)(*&"#S*&*(&*0#/-&(1)&'#I/.#%&:,*&$-#&#74(1#-#'&*(&-/%#&,4&$5$/)%*&*0(%#&10(&%,-I#/"#'&*0#7&/)&*0#&%0$'(1%X&#).(,-$5#'&*(&*#""& *0#/-& %*(-/#%X& $)'& :#& .(74#)%$*#'& S(-& *0#& $:,%#& .(,"'& :#& )(*& ()"M&-#%*(-$*/I#&:,*&*-$)%S(-7$*/I#<&&

&

&143 See Sunstein, supra note 141, at 2033. 144 See ICRC Customary Rules, 2005, supra note 56, Rule 150. Reparation. 145 See Reisman, supra note 70, at 398.

OO& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

#/&&J012&*2;B69;K@2&$952@@3429A2&+9;@8B5L&.5;9F;<F&&

In introducing a new duty of care, a “reasonable intelligence analyst” standard one must be cognizant of a set of factors that should be assessed to determine the utility of such a duty. Well tested and proven doctrines of tort law already establish a traditional multi-pronged assessment which includes such factors as: (1) the foreseeability of harm to potential victims; (2) the closeness of the connection between the perpetrators conduct and the injuries suffered; (3) the moral blame attached to the perpetrators’ conduct; (4) the policy rationales of preventing future harms; (5) the extent of the burden on the perpetrator and the consequences to the community of imposing a duty to exercise extra duties of care with resulting liability for breach; (6) the availability, cost, and prevalence of insurance for the risk involved.%")

This is a social policy analysis which takes into account the notion that punishing people or States for “conduct they cannot avoid at reasonable cost will have either no effect or a bad (inefficient) effect.”*%"! Applying the factors in our case will suggest, however, that a duty of care is necessary and that the risks are avoidable subject only to the adoption of certain basic principles. The adoption of these practices, holding intelligence officers to a basic duty to act reasonably, and then monitoring continuously that they abide by this duty, is far from costly to the intelligence community and will result in a reduction in risk without hindering on the ability of the State to fight legitimate wars or address real threats to the security of forces and the nation. We know this to be true, because many of the duties described below are already implemented by most intelligence agencies around the world today. Putting a legal gloss over these principles will only enhance their normative value and their ability to play an even more effective role in deterring bad conduct.

Moreover, there is no insurance policy for the victims of these aerial strikes based on faulty intelligence assessment. Most victims live in areas of the world where insurance is scarce and expensive. Even those lucky few with life insurance, for example, will quickly find a “wartime” based exclusion denying them coverage. Guaranteeing their recovery in the case of a reprehensible loss is one of the functions of our international system. As the UN General Assembly recognized: “victims should be treated with compassion and respect for their dignity, have their right to access to justice and redress mechanisms fully respected, and that the establishment,

&!$&$!""B#"J2J=$b,<<,'A$?'&$7#$>'/1+..@$\T[$>#:*JJ#\5$PSR[e$K1F1-0&$.($6-,7#$.($C'<#$

7#$ :*J1",."$ C.*"0=$ [$ C'<#$ Q0;$ _S`=$ _RZ=$ R\S$ C'<#KJ0"#\5$ [PQ=$ [P\$ 2#\5$ _Q_$ ORSPZUe$K.3<'-5$7#$C;",&0,'-=$_T$C'<#$R5$PSZ$PP\=$`S$C'<#KJ0"#$T`=$[[\$2#R5$Q_P$OPT_ZU$

!$'$K+/-!"2&%#&23.*#"=&N-#&k/3*3)+/.&34&I(.$+/#$RRQ$OPTZPU#$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& OW&

strengthening and expansion of national funds for compensation to victims should be encouraged, together with the expeditious development of appropriate rights and remedies for victims.”%"+

The following pages begin to outline the scope and nature of this duty of care. This is only an initial sketch. Ultimately, intelligence officers, military commanders and lawyers, human rights activists, and international organizations should come together to begin a more robust dialogue over this duty. My hope is that the following preliminary articulation could nonetheless assist by offering a starting point for this long overdue conversation.

A< =&8()*#f*3T$%#'&^,*M&(S&8$-#&

“Perfect information”, whatever this term might mean, is not an attainable

goal in times of war. That is not to say, however, that there are no objective categorizations of appropriate and inappropriate methodologies for intelligence processing, analysis, and verification. Nor does it mean that we must accept all intelligence faults and errors in war as inevitable and thereby legally inconsequential.

Instead, we should recognize that while partial information is the coin of the realm, qualitatively more accurate and detailed information should always be sought, and on many occasions may be achieved. The threshold of evidence necessary for launching an aerial operation should be defined and examined by lawyers, intelligence professionals, and military commanders.

In developing such a standard, it is important to clarify that no guideline could be systematically applied to all wartime intelligence operations. Surely, the verifications required may vary on the basis of context: the military, intelligence, and technological capacities of the targeting State, and the nature of the environment in which the operation is launched. The International Commission of the Red Cross (ICRC) clarified this point:

“Obviously, the standard of doubt applicable to targeting decisions cannot be compared to the strict standard of doubt applicable in criminal proceedings but rather must reflect the level of certainty that can reasonably be achieved in the circumstances. In practice this determination will have to take into account, inter alia, the intelligence available to the decision maker, the urgency of the situation, and the harm likely to result to the operating forces or to

&!$($ ?'&,/$ 2",-/,J<1&$ '-5$ M*,51<,-1&$ .-$ 0;1$ K,F;0$ 0.$ '$ K1A158$ '-5$ K1J'"'0,.-$ (."$

E,/0,A&$.($M".&&$E,.<'0,.-&$.($4-01"-'0,.-'<$D*A'-$K,F;0&$)'3$'-5$:1",.*&$E,.<'0,.-&$.($4-01"-'0,.-'<$D*A'-,0'",'-$)'3$$%5.J015$'-5$J"./<',A15$+8$M1-1"'<$%&&1A+<8$"1&.<*0,.-$_SWP[`$.($P_$!1/1A+1"$RSSQ#$

OQ& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

persons and objected protected against direct attack from an erroneous decision”.149

On the issue of urgency, as raised by the ICRC, the Israeli Government

makes an accurate observation:

“[D]uring ground operations, fire from a building near an infantry platoon may demand an immediate response, and the platoon may not have access to real-time data regarding the presence of civilians or the nature of surrounding structures. In such exigent circumstances, the platoon will have to rely on whatever partial information it does have, in addition to its prior training on the Law of Armed Conflict. The legality of the platoon’s conduct must be assessed in light of what a reasonable commander would or would not have done under the same or similar circumstances.” 150

In the context of aerial strikes we can easily think of situations where

urgency might be invoked. Think, for example, of opportunity strikes, such as a missile launcher coming out of the ground or a known terrorist operative coming out of hiding. They leave militaries with a limited window to collect intelligence and launch an attack. However, as a whole, aerial strikes often are planned days, weeks, and sometimes months in advance. For these kinds of cases, where the launching is decided in vacuum-like conditions, certainly the modest regulations I suggest below are of particular relevance.151 >< L*-,.*,-$"&$)'&!-(.#',-$"&^,*/#%&

B0#& "$%*& '#.$'#& 0$%& %##)& $& -/%#& /)& *0#& %.0("$-"M& -#%#$-.0& (S& 5"(:$"&$'7/)/%*-$*/I#&"$1<&=*&/*%&-((*X&*0#&4-(]#.*&*,-)%&*(&*0#&5#)#-$"&4-/)./4"#%&*0$*&$-#&.(77()&*(&7(%*&'(7#%*/.&$'7/)/%*-$*/I#&4-(.#%%#%X&$)'&7$2#%&*0#&."$/7&*0$*&*0#%#&4-/)./4"#%&$-#&$"%(&$&:-($'#-&S#$*,-#&/)&*-$)%35(I#-)7#)*$"&-#5,"$*/()%<152&=)M&4-(4(%$"&*(&-#/)*#-4-#*&*0#&)(-7$*/I#&"$)'%.$4#&(S&CEF&

&!$)$ INTERNATIONAL COMMISSION OF THE RED CROSS, INTEROPERATIVE GUIDANCE ON

THE NOTION OF DIRECT PARTICIPATION IN HOSTILITIES UNDER INTERNATIONAL HUMANITARIAN LAW 76 (Nils Meltzer ed., 2009).$

!%*$!""$4&"'1<$RSP[$M'X'$C.-(<,/0$K1J."0=$%=?&'#-.01$RZ=$'0$q$\R`#$!%!$!""#'(%/#4CKC$kmJ1"0$9110,-F$K1J."0=$%=?&'#-.01$RR=$'0$QT$O5,&0,-F*,&;,-F$+10311-$

J"1YJ<'--15$ .J1"'0,.-&$ '-5$ d58-'A,/$ 0'"F10,-Ff=$ -.0,-F$ 0;'0$ (."$ 0;1$ <'001"$ /'01F."8$ .($.J1"'0,.-&$dJ".J."0,.-'<,08$'&&1&&A1-0&$'"1$'<&.$/'"",15$.*0$+*0$,-$'$/.AJ"1&&15$0,A1("'A1=$3,0;$<1&&$,-01<<,F1-/1$'-5$0;1"1(."1$-.0$-1/1&&'",<8$0;".*F;$0;1$&'A1$J"./1&&1&#fU#$

!%"$>."$(*"0;1"$"1'5,-F$&11=$1#F#=$?1-5,/0$B,-F&+*"8$10$'<#=$<4"#3+"&2"1A"#/F#5(/9'(#*-+)1)%.&'.)L"#G'I=$_Z$)!5&a&C3*$#)0#&2"39.#$PQ$ORSSQU$OdF<.+'<$'5A,-,&0"'0,71$<'3$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& O[&

:M&/)*-(',./)5&)#1&',*/#%&(S&.$-#&$)'&%*$)'$-'%&(S&"/$:/"/*M&7,%*&-#.(5)/`#&*0$*&/*%&*$-5#*&$,'/#).#&/%&*0#&/)*#""/5#).#&$5#)./#%&*0#7%#"I#%<&C*&/%&$:(,*&*0#/-&/).#)*/I#&%*-,.*,-#&$)'&*0#/-&/)*#-)$"&4-(.#',-#%<&P/I#)&*0$*&*0#%#&$-#&%*$*#& $5#)./#%X& %4#$2/)5& /)& *0#& "$)5,$5#& (S& $'7/)/%*-$*/I#& "$1& $)'& /*%&%*-,.*,-$"& %$S#5,$-'%& (S& 4$-*/./4$*/()X& *-$)%4$-#).MX& -#$%()& 5/I/)5X& $)'&(I#-%/50*X&%##7%&4$-*/.,"$-"M&$4*<&=)M&$%%#%%7#)*&(S&$&:-#$.0&(S&*0#&',*M&(S&.$-#&1(,"'&*0,%&:#&:$%#'&()&$&S$/",-#&*(&.(74"M&1/*0&:$%/.&4-(.#',-$"&.0#.24(/)*%<&&

(a) Duty of Corroboration &U(%*& (S& (,-& 0/%*(-/.$"& .$%#& %*,'/#%& /)'/.$*#& *0$*& S$,"*%& (..,-&

4-#'(7/)$*#"M& ',#& *(& $)& (I#-3-#"/$).#& ()& $& %/)5"#& 4(/)*& (S& '$*$<& B0/%&/).",'#%&/)$..,-$*#&7$4%X&/).().",%/I#&'-()#&S((*$5#X&(-&$&%/)5"#&*/4&S-(7&$)& /)S(-7$)*& (-& *0/-'34$-*M& /)*#""/5#).#& $5#).M<& In the Chinese embassy case, the U.S. recognized that part of the fault lied with its intelligence agencies’ “susceptibility to a single point of database failure.”%,( Reliance on a single source is often appealing to intelligence analysts. It follows from the natural inclination towards the lex parsimoniae of Occam’s razor. Multiple sources lead to conflicting evidence and paradoxes which resolution might require careful and painstaking work.

The only way to break away from potential over-confidence and selection bias is through deliberative and participatory procedures. Such procedures should welcome ingenuity and creativity and encourage members to challenge traditional conceptions. In certain cases (perhaps in most cases of aerial strikes), it might even be called for, to have a group of analysts involved in counterfactual reasoning, challenging every single one of the agency’s assessment prepositions in the lead up to the publication of that assessment or the launching of an attack. Adopting such :$%/.& -#+,/-#7#)*%& (S&I#-/S/.$*/()& /%& *0#&()"M& %,-#&1$M& *(&.0$""#)5#&4("/*/./`#'& /)*#""/5#).#&$)'&5-(,4*0/)2&7#)*$"/*/#%X&%(&4-#I$"#)*&/)&7/"/*$-M&/)*#""/5#).#&(4#-$*/()%<&

K#"$*#'"MX&1#&0$I#&1/*)#%%#'&0(1&*0#&/)*-(',.*/()&(S&)#1&%,-I#/""$).#&*(("%&0$I#&*-/55#-#'&S$"%#&.()S/'#).#&/)&*0#%#&-/%/)5&b*#.0)("(5M&5('%c&$%&$)& $"*#-)$*/I#& *(& $& 7(-#& .$-#S,"& $%%#%%7#)*& (S& *0#& S$.*%. I have already

&g,&$ 51(,-15h$ '&$ /.AJ",&,-F$ 0;1$ A1/;'-,&A&=$ J",-/,J<1&=$ J"'/0,/1&$ '-5$ &*JJ."0,-F$ &./,'<$*-51"&0'-5,-F&$0;'0$J".A.01$."$.0;1"3,&1$'((1/0$0;1$'//.*-0'+,<,08$.($F<.+'<$'5A,-,&0"'0,71$+.5,1&=$ ,-$ J'"0,/*<'"$ +8$ 1-&*",-F$ 0;18$ A110$ '51c*'01$ &0'-5'"5&$ .($ 0"'-&J'"1-/8=$J'"0,/,J'0,.-=$ "1'&.-15$51/,&,.-=$ '-5$ <1F'<,08=$ '-5$+8$J".7,5,-F$ 1((1/0,71$ "17,13$.($ 0;1$"*<1&$'-5$51/,&,.-&$0;18$A'@1fUe$K,/;'"5$?#$:013'"0=$6J!J#*-+)1)%.&'.)L"#G'I:#*#8/-"(#F/&#5(/9'(#*-+)1)%.&'.)L"#G'I=$_Z$)!5&a&C3*$#)0#&2"39.#$_\$ORSSQU#$

!%#$!""#%=?&'#-.01$PPP$'-5$'//.AJ'-8,-F$01m0#$

Od& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

alluded to a potential over-reliance on technology, namely on visual intelligence (VISINT) in the context of surveillance drones, as a primary (and on occasion sole) source of intelligence for targeting verification purposes. The Intercept leaking of the Drone Papers (Pentagon’s top-secret reports concerning Obama-era policies on targeting) further show the “critical shortfalls” in over-reliance on signal intelligence (SIGINT), namely location tracking on the basis of GSM mobile networks, when engaging targets.%,"

This is exactly the cue for law and technology to sidle onto the dance floor and begin their "choreographed exchange".155 For some, however, international law no longer regulates the development and use of new technologies; rather, the tail is wagging the dog, with technology becoming “an invisible hand guiding, destroying, and creating international law”.156 The fast pace advancement of intelligence capabilities, as part of the broader contemporary cyber and information warfare, only proves the inadequacy of regulation as a reliable panacea. As one commentator had suggested:

“The development of advanced information warfare technologies and techniques and the continuing global diffusion of information technology illustrate the fluidity of the world that law attempts to govern. No law can change, as swiftly as can technology; unless law is to somehow stop technology's seemingly inexorable worldwide progress, it cannot fully control the use of its fruits for warfare. Legal measures can thus supplement, but not supplant, vigilance, preparedness, and ingenuity.”157

The question then becomes should we accept as inevitable some form of

soft technological determinism, and what would that entail to our ability to govern modern forms of wartime intelligence collection. For the purposes of this paper suffice it to say that any reasonable intelligence analyst must take into account in their assessments the limits of the technologies they employ.

&!%$$!""#C."'$C*"",1"$'-5$2101"$9''&&=$7)&)12#D()1-:#7('I"-#$1."(()2"1A"#'1-#.4"#G)+).%#

/F#W&/1"# <"A41/(/2,=$ N-#& 4*$#"/#0$$ Oi/0#$ PQ=$ RSPQU=$ ;00J&VWW0;1,-01"/1J0#/.AW5".-1YJ'J1"&W(,",-FY+<,-5W#$

155 Timothy R. Coughlin, The Future of Robotic Weaponry and the Law of Armed Conflict: Irreconcilable Differences? 17 UCL JURIS. REV. 67, 67-68 (2011) (“The dichotomy between law and technology has long been tenuous in the best of times, and irreconcilable at the worst. Answering to different masters, technological development and legal structures are in a constant state of ebb and flow, with each pushing the contours of the other in choreographed exchange of concessions and compromises.”).

156 Colin B. Picker, A View from 40,000 Feet: International Law and the Invisible Hand of Technology, 23(1) CARDOZO L. REV. 149, 202 (2002).

157 LAWRENCE T. GREENBERG, SEYMOUR E. GOODMAN & KEVIN J. SOO HOO, INFORMATION WARFARE AND INTERNATIONAL LAW 37 (1998).

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& OJ&

For example, in its contemporary form aerial surveillance or signal intelligence cannot serve as the sole basis for verification and justification of attacks. Each of these sources must be corroborated with other intelligence sources before an operation is authorized. Establishing as unreasonable a decision to launch an attack only on the basis of one such source, begins to draw the legal lines in assigning liability. At the same time, any standard developed must evolve in light of ongoing changes in the technology. It could very well be that future rule prescribers and rule appliers will view new advancements in remote sensing, hacking, and communications interception as ones that necessitate further line erasure and new line drawing for our duty of care.

(b) Duty of Documentation &H,-&0/%*(-/.$"&.$%#&%*,'/#%&/)'/.$*#&*0$*&*0#&()"M&1$M&*(&'#*#-7/)#X&2I&

76B5&D;A5ECX&/S&$&:-#$.0&(S&$&',*M&(S&.$-#&0$'&(..,--#'&/%&/S&*0#-#&/%&$&I#-/S/#'&*-$/"& (S& #I/'#).#& *(& -#I/#1<& Zf$7/)$*/()& (S& $""& *0#& /)S(-7$*/()& *0$*& 0$'&*-$)%4/-#'&()&*0#&:$%/%&(S&10/.0&$&'#./%/()&*(&#)5$5#&/)&$)&$**$.2&1$%&7$'#X&/%&$:%(",*#"M&.-,./$"&/)&-#$.0/)5&$&.().",%/()&$%&*(&*0#&-#$%()$:"#)#%%&(S&*0#&(4#-$*/()<&B0/%& -#+,/-#%& *0#& $'(4*/()&(S& %*-/.*& $)'&1#""3'#S/)#'& -,"#%& S(-&4-#%#-I/)5&*0#&-#.(-'&(S&*0#&I$-/(,%&$.*/()%&$)'&'#./%/()%&*0$*&1#-#&7$'#&',-/)5&*0#&(4#-$*/()&$)'&*0-(,50(,*&*0#&/)*#""/5#).#&.M."#<&C*&S,-*0#-&#)*$/"%&*-$/)/)5& /)*#""/5#).#&(SS/.#-%& $:(,*& -,"#%& -#"$*/)5& *(& .0$/)&(S& .,%*('M&$)'&*0#/-&(:"/5$*/()%&*(&4-#%#-I#&*0#&'$*$&:(*0&:#S(-#&$)'&$S*#-&#$.0&(4#-$*/()<&

C).-#$%#'& /)S(-7$*/()& /%& *(& ()"M& 5,$-$)*##& S(-& -#$.0/)5& #SS/./#)*&'#./%/()%& $%& *(& *0#& .(%*3:#)#S/*& $)$"M%/%& (S& #f*#)'/)5& *0#& %*$)'$-'& (S&-#$%()$:"#)#%%&*(&)#1&%/*,$*/()%&$%&*0#M&$-/%#<& &&

(c) Duty of Transparency &

It is important to acknowledge that “secrecy is not per se unlawful,”158 but it does have the potential of abuse,159 given the inclination towards

&158 W. MICHAEL REISMAN &. JAMES E. BAKER, REGULATING COVERT ACTION:

PRACTICES, CONTEXTS, AND POLICIES OF COVERT COERCION IN INTERNATIONAL AND AMERICAN LAW 141 (1992).

159 Hans Born & Aidan Wills, Beyond the Oxymoron: Exploring Ethics through the Intelligence Cycle, in 2 ETHICS OF SPYING: A READER FOR THE INTELLIGENCE PROFESSIONAL 52, 54 (Jan Goldman ed., 2010) (“Ethical standards are of particular significance in the activities of intelligence services due to the nature of work they conduct and the veil of secrecy under which much of their work takes place. Furthermore, in even the most transparent democratic systems, it is feared that a proportion of the work of intelligence services falls beyond the reach of both legislation and oversight bodies.”).

Wa& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

corruption of closed processes.160 As noted by David Pozen in the “absence of perfect virtue, secrecy creates greater opportunities for officials to pursue personal or partisan gain, to engage in logrolling or horse trading, and to commit legal and ethical abuses... Secrecy exacerbates the principal-agent problem inherent in representative democracy and opens the door to tyranny.”161

What is even more worrying, however, from an international legal perspective, is the devastating blow delivered by the secrecy of the intelligence profession to the international constitutive process. Lacking a codifying treaty, international legal regulation is dependent upon state practice to effectively grow. Unfortunately, secrecy inherently prevents custom from evolving. Professor Ratner has defined it best: &

“The challenge to international lawyers posed by intelligence gathering is somewhat daunting: How can international norms, processes, and institutions possibly play a role in regulating an activity that by its very nature is so secret that states deliberately reveal very little about how they carry it out?”162 &&

Operating without the compass of international custom, intrepid legal advisors are forced to navigate a treacherous international legal terrain and provide their States’ with unsubstantiated answers to policy questions, which could have horrendous consequences. Even more broadly, there is room to also challenge the efficacy of “security through secrecy” doctrines both from a rule-of-law prism,163 and as a practical matter in a flattening world where everyone owns a smartphone and social media reigns supreme.164&B0#&%$7#&.$)&:#&%$/'&S-(7&$)&$'7/)/%*-$*/I#&"$1&4#-%4#.*/I#X&$5#)./#%&$-#&,)'#-%*(('&*(&$.*&'/SS#-#)*"M&10#)&*0#/-&4("/.M7$2/)5&(..,-%&/)&*0#&"/50*X&-$*0#-&*0$)&/)&%#.-#*<%),&

&160 See generally, Clare Birchall, Introduction to ‘Secrecy and Transparency’: The

politics of Opacity and Openness 28 THEORY CULTURE & SOC. 7, 11 (2011). 161 David E. Pozen, Deep Secrecy, 62(2) STAN. L. REV. 257, 278 (2010). 162 Steven R. Ratner, Introduction, 28 MICH. J. INT'L L. 539, 539 (2007). 163 Maure L. Goldschmidt, Publicity, Privacy and Secrecy, 7(3) WEST. POL. Q. 401, 410

(1954) (“The efforts to defend national security through secrecy are, in part at least, self-defeating. If everything is classified, classification soon loses its significance for those who handle secret documents... Even more important are the effects of secrecy upon public opinion. The lack of reliable information increases fear and uncertainty, and encourages rumor and the making of reckless and irresponsible charges”).

164 Reisman & Baker, supra note 158, at 142 (“In contemplating any covert operation… assume that it will become public knowledge much sooner than you would like… The illusion of secrecy can provide a false and treacherous sense of security.”).

!&%$!""B#"J2J=$).*,&$?"'-51,&=$i0;1"$21.J<1n&$9.-18$_R$OPTT\U$OdJ*+<,/,08$ ,&$ G*&0<8$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& WA&

E(1#I#-X& $%& CEF& /%& .,--#)*"M& %*-,.*,-#'& *0#-#& /%& )(& (:"/5$*/()& *(&4,:"/./`#& /)S(-7$*/()& .().#-)/)5& *0#& 1$M& /)*#""/5#).#& 1$%& .(""#.*#'X&4-(.#%%#'X& $)$"M`#'X& I#-/S/#'X& (-& ,*/"/`#'<%))& H:I/(,%"MX& /)& *0#& )$7#& (S&%(,-.#& 4-(*#.*/()X& )(*& $""& /)S(-7$*/()& 7$M& :#& 7$'#& 4,:"/.& $*& $""& */7#%<&e()#*0#"#%%X&$)M&"#5$"&S-$7#1(-2&%,--(,)'/)5&CEF6%&/)*#""/5#).#&S,).*/()%&7,%*&.$""&()&L*$*#%&*(&'/%."(%#&5-#$*#-&/)S(-7$*/()&$)'&4-(I/'#&7(-#&-(:,%*&'#*$/"%& ()& *0#& 1$M& *0#/-& /)*#""/5#).#& .M."#X& 10/.0& 1$%& /)0#-#)*& /)& *0#&'#./%/()& *(& "$,).0& #$.0& *$-5#*#'& %*-/2#X& 1$%& .()',.*#'<& CEF& 7/50*& $"%(&:#)#S/*&S-(7&/).#)*/I/`/)5&C)*#""/5#).#&$5#)./#%&*(&#%*$:"/%0&7#.0$)/%7%&S(-&10/%*"#:"(1#-&4-(*#.*/()%X& /)&(-'#-& *(& /).-#$%#&(44(-*,)/*/#%& S(-& S$,"*%& *(&.(7#& /)*(& *0#& "/50*& $)'& :#& %.-,*/)/`#'& $)'& $''-#%%#'& $0#$'& (S& $)& #--(-&7$*#-/$"/`/)5<&&

C*&/%&/)&*0/%&.()*#f*X&*0$*&*0#&B-,74&$'7/)/%*-$*/()6%&'#./%/()&*(&-(""&:$.2&$)&H:$7$3#-$&4("/.M&()&'/%."(%,-#%&(S& %*$*/.$"&'$*$&()&./I/"/$)&.$%,$"*/#%&S-(7&$/-%*-/2#&*$2#%&,%&:$.21$-'%&$)'&)(*&S(-1$-'<167&=%&()#&.(77#)*$*(-&)(*#'& $#-/$"& %*-/2#%& *$2#& 4"$.#& /)& $)& $"-#$'M& *-(,:"/)5& #)I/-()7#)*& (S&#)0$).#'&%#.-#.MX&b10#-#&*0#&5(I#-)7#)*X&)#1%&(-5$)/`$*/()%&$)'&0,7$)&-/50*%&5-(,4%&0$I#&"/7/*#'&I/%/:/"/*MX&$)'&*0#&#)#7M&0$%&%0(1)&$&4-(4#)%/*M&*(&$**#74*&*(&7$)/4,"$*#&*0#&/)S(-7$*/()<c168&

In addition, we should eliminate jurisdictional hindrances on the ability to pursue further public discourse around targeting decisions. Consider in this regard the Jaber case. On August 29th, 2012 an American drone strike in Yemen killed Ahmed Salem bin Ali Jaber, an imam who preached against al-Qaeda's ideology. Salem was not the target, but a collateral damage. That attack was aimed at 3 al-Qaida members which Salem was meeting in an attempt to try and change their views. Faisal bin Ali Jaber, one of Salem's family members, received 100,000$ as condolence payments from the U.S.

&/.AA1-515$'&$'$"1A158$(."$&./,'<$'-5$,-5*&0",'<$5,&1'&1&#$:*-<,F;0$,&$&',5$0.$+1$0;1$+1&0$5,&,-(1/0'-0&e$1<1/0",/$<,F;0$0;1$A.&0$1((,/,1-0$J.<,/1A1-fUe$9#)3"!*2()&43"&$-#&D#!2.&

34& k8#/($+1#& !#0!"$)#*$.& !*2& %,#*/+#.& 3*& N"!*.0!"#*/6& !*2& i0#*&

M31#"*)#*$=&i44+/#&34&$-#&2"#.+2#*$$OI'-#$RP=$RSSTU$O4-$;,&$(,"&0$J"1&,51-0,'<$5,"1/0,71$*J.-$ 0'@,-F$ .((,/1$ 2"1&,51-0$ i+'A'$ ;'5$ -.015V$ dM.71"-A1-0$ &;.*<5$ +1$ 0"'-&J'"1-0#$N"'-&J'"1-/8$J".A.01&$ '//.*-0'+,<,08$ '-5$J".7,51&$ ,-(."A'0,.-$ (."$ /,0,X1-&$ '+.*0$3;'0$0;1,"$M.71"-A1-0$ ,&$ 5.,-F#fUe$!""# '(%/$B,-F&+*"8=$ %=?&'#-.01$ PQR=$ '0$ \Z$ Od!1/,&,.-'<$0"'-&J'"1-/8$'-5$'//1&&$0.$,-(."A'0,.-$'"1$,AJ."0'-0$(.*-5'0,.-&$(."$0;1$1((1/0,71$1m1"/,&1$.($J'"0,/,J'0,.-$",F;0&$'-5$",F;0&$.($"17,13#$N;18$'<&.$J".A.01$'//.*-0'+,<,08$5,"1/0<8$+8$1mJ.&,-F$'5A,-,&0"'0,71$51/,&,.-&$'-5$"1<17'-0$5./*A1-0&$0.$J*+<,/$'-5$J11"$&/"*0,-8fU#$

!&&$!""=$4&"'1<$RSP[$M'X'$C.-(<,/0$K1J."0=$%=?&'#-.01$RZ=$'0$q$RZ_$O(-#$[\ZU#$167 Charlie Savage, Trump Revokes Obama-Era Rule on Disclosing Civilian Casualties

From U.S. Airstrikes Outside War Zones, N.Y. TIMES (Mar. 6, 2019), https://www.nytimes.com/2019/03/06/us/politics/trump-civilian-casualties-rule-revoked.html.

168 Id.

W>& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

but no official apology nor recognition of fault was ever provided by the Government. So he brought a federal lawsuit. On Friday 30, 2017, the United States Court of Appeals for the D.C. Circuit issued its much-anticipated judgment. The Court dismissed the case as a nonjusticiable political question. The Court argued that "it is the Executive, and not a panel of the D.C. Circuit, who commands our armed forces and determines our nation’s foreign policy".169

But the Court was never called to meddle in the way the President commands the armed forces or runs his foreign policy. Rather, the Court was asked to acknowledge that modern-day premeditated remote drone strikes involve substantive legal discourse and legal interpretation. This kind of discourse and interpretation is in fact deeply within the "wheelhouse of the judiciary"170 as the D.C. Circuit termed it. Courts are thus well-equipped, not ill-equipped, to engage in these analyses and to assess the legal standards adopted by the executive prior to launching such strikes. They also offer an important venue through which to fight the phenomenon of groupthink. If intelligence officers and military commanders knew that their pre-targeting decisions could be the subject of potential public litigation later down the line, they might be more cautious in providing detailed reasoning during deliberations and ahead of the authorization of strikes.171

,/&&M355394&512&%2=&(E58&6D&,;<2&:35139&-I3B5394&$N!&

The rules of state responsibility for internationally wrongful acts, “that is to say, the general conditions under international law for the state to be considered responsible for wrongful actions or omissions, and the legal consequences which flow there-from,”172 are secondary rules. As such, for them to be invoked, it is required that we first identify a breach of a primary international obligation.

However, as exemplified in the ICTY’s Expert Committee Report, and as currently interpreted an attack that is indiscriminate, disproportionate, or otherwise in violation of the laws of war, would not be considered “a wrongful act” if it was conducted on the basis of faulty intelligence. In other words, if ex ante, at the time the operation was authorized, the information

&169 Ahmed Salem Bin Ali Jaber v. United States, D.C. Cir. No. 16-5093, 15 (Jun. 30,

2017). 170 Id., at 13. 171 For further reading on the deterrence effect of litigation on the wartime decision

making of politicians, military commanders, and rank and file soldiers see Haim Abraham, Combatant Activity, Tort Liability, and the Question of Over-Deterrence (draft on file with author).

!'"$ ILC Articles on State Responsibility for Internationally Wrongful Acts with commentaries, U.N. Doc. A/56/10, 31 (2001)#$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& WD&

provided to policy makers made the operation lawful under the laws of war, then an ex post revelation of intelligence omissions will not invalidate the operation’s legality. Here lies the accountability gap. This allows States to revert to “a fog of war” accident claim, which will excuse them of liability whenever their targeting operations result in unanticipated-at-the-time civilian casualties.

Therefore, a normative shift must take place, whereby we recognize that certain intelligence omissions and faults can be prevented and are not inevitable. If those faults, for reasons of a lack of due diligence, or recklessness, or willful blindness, are not corrected and result in loss of civilian life, such loss constitutes an unreasonable mistake that will result in a violation of API Art. 57. It would mean, that compensation would be provided and that public expressions of regret would be issued as a matter of a legal obligation, not moral decorum.

By rejecting the ICTY’s Expert Committee’s Report, we avoid the need for an unrealistic amendment of the Geneva Conventions. Instead, we can begin the process of extending the interpretation of the principle of precautions in so to more closely align with the objectives of IHL to mitigate and reduce the harmful consequences of war.%!(

To accommodate this shift we may wish to explore what Rebecca Crootof calls “informal lawmaking.” This process involves such sources as “Common understandings based on international and transnational dialogue, nonbinding resolutions and declarations, professional guidelines and codes of conduct, civil society reports and policy briefs, industry practice, and even domestic laws and policies. such sources are likely to be both more narrowly tailored and more flexible than treaty provisions, and thus better able to address unanticipated technological breakthroughs.”%!"

An ICRC interpretive guidance on Article 57, agenda-setting sessions and

&!'#$!""#K1+1//'$C"..0.(=$<4"#0)(("&#C/9/.%#*&"#>"&":#G"2'(#'1-#O/()A,#$+?()A'.)/1%=$

\_$C!"23%3&)#&K#1#$PZ\`=$PTSPYTSR$ORSPQU#$!'$$%&$ J'"0$ .($ 0;,&$ J'J1"$ 4$ 1m/<*515$ '-8$ '-'<8&,&$ .($ ,-5,7,5*'<$ <,'+,<,08$ (."$ &J1/,(,/$

,-01<<,F1-/1$.((,/1"&#$6-51"$1m,&0,-F$ ,-01"-'0,.-'<$/",A,-'<$ <'3=$ ,-5,7,5*'<&$A'8$+1$;1<5$/",A,-'<<8$<,'+<1$.-<8$(."$;'"A&$"1&*<0,-F$(".A$3,<<(*<$/.-5*/0$'-5$"1/@<1&&$'/0,.-=$0;'0$,&$d0;1$ '00,0*51$ .($ '-$ 'F1-0$3;.=$3,0;.*0$ +1,-F$ /1"0',-$ .($ '$ J'"0,/*<'"$ "1&*<0=$ '//1J0&$ 0;1$J.&&,+,<,08$.($,0$;'JJ1-,-F#f$O!""#4*$n'&C3))#&34&$-#&K#2&C"3..=&C3))#*$!"6&3*&$-#&

%22+$+3*!'&2"3$3/3'.&34&Z&I(*#&PT``&$3&$-#&M#*#1!&C3*1#*$+3*.&34&PR%(,(.$&

PT[T=&!$&TT[& OH71&$:'-5.X=$C;",&0.J;1$:3,-'"&@,$a$?"*-.$j,AA1"A'--$15&#=$PTZ`U=$;00JVWW333#<./#F.7W""W("5W9,<,0'"8r)'3WJ5(WC.AA1-0'"8rMCr2".0./.<&#J5(e$ %""# '(%/$2".&1/*0."$7#$!1<'<,s=$C'&1$L.#$4NYT_YRPYN=$N",'<$C;'A+1"$I*5FA1-0=$qq$[\`=$[\T$O4-0n<$C",A#$ N",+#$ (."$ 0;1$ >."A1"$ H*F.&<'7,'$ L.7#$ P_=$ PTTZU=$;00JVWW333#,/08#."FWmW/'&1&WA*/,/W0G*FW1-WTZPPP_rG*5Fr1-#J5(#U#$ N;,&$ 3.*<5$ &11A$ 0.$1m/<*51$A.&0$.($0;1$('*<08$,-01<<,F1-/1$.J1"'0,.-&$5,&/*&&15$,-$0;1$/'&1$&0*5,1&$'-5$0;1"1(."1$&11A&$0.$+1$.*0&,51$0;1$&/.J1$.($.*"$'-'<8&,&#$

WO& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

reports by special procedures at the Human Rights Council in Geneva, and public-facing state military guidelines and codes of conduct for wartime intelligence production are but some examples for what an informal lawmaking program could look like. &

8*,&'."/*,&

In the ancient world war-chiefs relied on mysticism in developing their grand military plans. These officeholders sought the help of oracles and soothsayers, calling on them to predict the outcomes of tactical and strategic military maneuvers and operations.175 The oracles “often hedged their bets by giving obscure and even deliberately ambiguous answers.”176 The Pythia in the Temple of Apollo at Delphi, for example, the most prestigious and authoritative oracle amongst the Greeks,177 was particularly known for providing her followers with polysemic prose.178 After bathing in and drinking from the waters of the Castalian spring, the Pythia burned bay fronds

&!'%$!""#B#+$-&k''+.=&2"#2+/$+3*&!*2&2"30-#/6&P`$OPT`\U$Od4-$0;1$'-/,1-0$3."<5=$

F1-1"'<&$'-5$J.<,0,/,'-&$/.-&*<015$."'/<1&$'-5$&..0;&'81"&$'&$"1'5,<8$'&$.*"$.3-$<1'51"&$/'<<$,-$01/;-,/'<$1mJ1"0&#fU#$$

!'&$$-J=$'0$PT#$$!''$ !""# 2"1"&'((,=$ b+''+!)& I#& ?"3!2=& N-#& i"!/'#V& %*/+#*$& !#'0-+& !*2& $-#&

:/+#*/#&?#-+*2&+$.&)3.$&:#/"#$.$TYPS$ORSS_U$O?".'5$-.01&$;.3$0;1$i"'/<1$.($!1<J;,$'-5$;1"$&'/"15$G*",&5,/0,.-$.-$0;1$&<.J1&$.($9.*-0$2'"-'&&*&$,-$/1-0"'<$M"11/1$d,-&J,"15$'$<,+"'"8fV$dN;1$<,01"'0*"1$+1F'-$3,0;$D.A1"$,-$0;1$1,F;0$/1-0*"8$?Ck$'-5$'//*A*<'015$(."$031<71$;*-5"15$81'"&Y$0;1$J1",.5$.($"1/."515$;,&0."8$,-$3;,/;$0;1$i"'/<1$'57,&15$A*<0,0*51&$.($ J,<F",A&=$ "1J1'015<8=$ ,0$3'&$ &',5=$ +8$ (."101<<,-F$ 0;1$ (*0*"1$ '-5$ "171'<,-F$ 0;1$ ;,551-$J"1&1-0#$D1"$ &/."1&$ .($ '5A,"1"&$ O'-5$ '$ (13$510"'/0."&U$ ,-/<*515$D1".5.0*&=$k*",J,51&=$:.J;./<1&=$2<'0.=$%",&0.0<1=$2,-5'"=$t1-.J;.-=$!,.5."*&=$:0"'+.=$2<*0'"/;=$),78=$I*&0,-=$i7,5=$)*/'-=$'-5$C<1A1-0$.($%<1m'-5",'=$A1-$"'-F,-F$(".A$M"11@&$0.$K.A'-&$0.$1'"<8$C;",&0,'-&###$D1".5.0*&=$0;1$('0;1"$.($;,&0."8=$0.<5$;.3$0;1$i"'/<1n&$F*,5'-/1$;1<J15$<1'5$0;1$M"11@&$0.$7,/0."8$,-$0;1$21"&,'-$3'"&#fU#$

!'($21";'J&$0;1$A.&0$'A+,F*.*&$31"1$0;1$J".J;1/,1&$F,71-$+8$0;1$!1<J;,/$."'/<1$0.$0;1$%0;1-,'-&$*J.-$0;1$21"&,'-$B,-F=$t1"m1&=$,-7'&,.-$,-0.$M"11/1$.($,-$0;1$&J",-F$.($[ZS$?C#$%&$ &*AA'",X15$ ,-$ 0;1$3."@&$.($D1".5.0*&=$ '-5$ 0"'-&<'015$ <'01"V$d%0$%0;1-&$ 0;1"1$3'&$(,1"/1$51+'01$.71"$3;'0$/.*"&1$.($'/0,.-$ 0;1$/,08$&;.*<5$ 0'@1#$ 4-c*,"8$3'&$A'51$.($ 0;1$!1<J;,/$."'/<1=$3;,/;=$'//."5,-F$0.$D1".5.0*&=$'0$(,"&0$*"F15$0;1$%0;1-,'-&$0.$d(<11$0.$0;1$1-5&$.($0;1$1'"0;=f$+*0$0;1-$&.(01-15$,0&$0.-1$&.A13;'0$3;1-$'&@15$(."$'$A."1$/.A(."0,-F$"1J<8#$N;,&$&1/.-5$."'/*<'"$"1&J.-&1=$3;,<1$&0,<<$5,"1=$&J.@1$.($&'(108$0;'0$/.*<5$+1$(.*-5$+1;,-5$'$d3..51-$3'<<f$'-5$.($'$5,&'&0".*&$+'00<1$ 0;'0$3.*<5$0'@1$J<'/1$.(($ 0;1$-1'"+8$,&<'-5$.($:'<'A,&#$L.3$0;1$51+'01$'0$%0;1-&$/1-01"15$.-$0;1$A1'-,-F$.($0;,&$."'/<1V$:.A1$,-01"J"1015$0;1$d3..51-$3'<<f$0.$A1'-$0;1$3..51-$(1-/1$&*"".*-5,-F$0;1$'/".J.<,&=$.0;1"&$0;1$3..51-$&;,J&$.($%0;1-&n$-13<8$+*,<0$-'78e$3;,<1$ 0;1$71"&1$J"./<',A,-F$ 0;'0$dD.<8$:'<'A,&$3,<<$51&0".8$A.0;1"&n$&.-&f$"',&15$0;1$c*1&0,.-$.($3;10;1"$M"11@$."$21"&,'-$d&.-&f$3.*<5$+1$51&0".815#f$O&11$D#"323$(.V&i*&$-#&b!"&>3"&M"##7&>"##23)&Y&:#'#/$+3*.&

4"3)&$-#&D+.$3"+#.$P\`YP\Z$OI'A1&$K.AA$15#=$RSS\UU#$$$

>3@$)3>Ag& !"#$%&'&#!$%(&)*+,!-.& WW&

and barley on an alter, and then mounted her holy tripod. As the priestess ascended into her mantic state, gazing into a dish of Kassotis spring water and laurel leaves, she began uttering her prophecies.179 Some authors claim this traditional process was a mere cover-up for a sort of self-induced frenzy, the result of a combination of intoxicating fumes and potent gases that rose into the temple from cracks in its bedrock. These brought about a trancelike state where the oracle would moan and wail, and the temple’s priests and Hosioi (noble Delphians who attended the Pythia)180 would interpret her wild cries according to an agreed-upon code.181 In modern warfare, the demand for foresight has only increased.182 While generals no longer turn to oracles, they continue to search for divine certainty amidst the tormenting fog of war. It is in this context that intelligence analysts have become our contemporary prophets. Their chairs and desks are their tripods, their computers their barley leaves. Instead of clay dishes filled with spring water, they have satellite dishes, intercepted underwater cables, and other surveillance and reconnaissance devices. Like the Pythia of Delphi, they too reach a genuine state of ecstasy, not one induced by illegal substances (so we hope), but rather a spiritual elevation achieved by a series of small, seemingly meaningless cathartic moments of intelligence puzzle piecing. And much in the same way that the military chiefs of old accepted the prophecies of oracles as the spoken word of god, so do our modern commanders susceptible to embracing intelligence

&!')$!""#I3.#0-&>3*$#*"3.#=&N-#&!#'0-+/&i"!/'#V&4$.&K#.03*.#.&!*2&i0#"!$+3*.&

5+$-&!&C!$!'3,(#&34&K#03*.#.$RR[YRRQ$OPT`ZU#$!(*$$-#$'0$RPT#$!(!$!""$)#5+.&K+/-!"2&>!"*#''=&N-#&C('$.&34&$-#&M"##7&:$!$#.&PZT$OPTS`U$Od0;1$

280;.-1&&$'&/1-515$,-0.$0;1$0",J.5=$'-5=$(,<<15$3,0;$0;1$5,7,-1$'((<'0*&$3;,/;$'0$<1'&0$0;1$<'01"$'F1&$+1<,1715$0.$'&/1-5$,-$7'J.*"$(".A$'$(,&&*"1$,-$0;1$F".*-5$+*"&0$(."0;$,-0.$3,<5$*001"'-/1=$3;,/;$3'&$J".+'+<8$&.A1$@,-5$.($'"0,/*<'01$&J11/;=$'-5$3;,/;=$0;1$g>/%)/)h=$0;1$p;.<8$.-1&=n$3;.$3,0;$0;1$J".J;10$&'0$'".*-5$0;1$0",J.5=$@-13$31<<$;.3$0.$,-01"J"10#$4-$'<<$0;,&$31$ -115$ 0.$ &*&J1/0$ -.$ /;'"<'0'-,&A###$N;1$ 1/&0'&8$ .($ 0;1$280;.-1&&$3'&$ -.$ 5.*+0$J1"(1/0<8$ F1-*,-1=$ .(01-$ 1m;'*&0,-F$ '-5$ &.A10,A1&$ 5'-F1".*&e$ 0;1$ +1<,1($ ,-$ 0;1$ 5,7,-1$'((<'0*&=$+8$3;'0171"$A1'-&$,0$3'&$,-&0,<<15=$/.*<5$J".5*/1$'$71"8$J.31"(*<$-1*".0,/$*J.-$'$&*&/1J0,+<1$01AJ1"'A1-0=$1m/,015$+1(."1;'-5$+8$'$/.*"&1$.($A'-0,/$&0,A*<'-0&#fU#$!""#'(%/#?".'5=&%=?&'#-.01$P``=$'0$PYP\#$i0;1"&$"1G1/0$0;,&$-'""'0,71=$/<',A,-F$0;'0$0;1$280;,'$3'&$ ,-$ /.AJ<101$ /.-0".<$ .($ ;1"$ /;."5&$ '-5$ <,J&#$ %&$ 0;1$ F.5$ J1-10"'015$ ;1"$ &.*<$ '-5$,AJ"1F-'015$;1"$A,-5$3,0;$7,&,.-&$.($ 0;1$ (*0*"1=$ &;1$1mJ1",1-/15$-.$ ("1-X8$'-5$ &J.@1$/<1'"<8$ '-5$ (<*1-0<8=$ 5,"1/0<8$ '55"1&&,-F$ 0;.&1$ &11@,-F$ ;1"$ '57,/1$ 3,0;$ -.$ -115$ (."$A15,'0,.-$O!""#>.-01-".&1=$%=?&'#-.01$P`Z=$'0$PTQYPT`U#$

!("$ M"1F."8$ k<51"=$ $1."(()2"1A"# )1# ;'&:# $.# K'1# D"# W"A)%)L"=$ QSORU$ :$(2#& +*&

4*$#''+,#*/#$P\=$P\$ORSS_U$O(."/1$'-5$,0&$1AJ<.8A1-0$'"1$&,F-,(,/'-0$,-$5",7,-F$.*0/.A1&$,-$/.A+'0#$D.3171"=$,0$,&$.J1"'0,.-'<$'-5$0'/0,/'<$,-01<<,F1-/1=$-.0$-1/1&&'",<8$-*A+1"&=$01/;-.<.F8=$."$0'/0,/&=$0;'0$/'-$;'71$0;1$A.&0$51/,&,71$,AJ'/0$.-$;.3$(."/1&$'"1$1AJ<.815$'-5$;.3$&*//1&&$,&$'/;,1715$,-$3'"0,A1$.J1"'0,.-&#fU#$

WQ& !"#$%&'&#!$%(&)*+,!-.& ?>3@$)3>A&

memorandums, and their authors’ deductive reasoning, as a reflection of some form of celestial manticism.183

This paper aims to set the scene for a conversation about the need to regulate intelligence as a pre-requisite for the protection of the goals of IHL. Intelligence is a queer profession, in the sense that it does not “fit neatly into the dichotomy that exist between military action (the exercise of hard power) and diplomacy (the exercise of soft power).”184 It is a realm in which existing treaties and legal regimes are replaced by “the more “perverse” set of norms and rules which appear to govern intelligence activities.”185

So far, international law has taken an all-or nothing (mostly nothing) approach to the development of wartime intelligence standards. However, just as intelligence is queer, our conceptual regulation of it must be queer too, elastic, non-conforming, adjustable along a spectrum. A duty of care standard offers such regulation and indeed may be easily embedded into our existing IHL frameworks. This paper should thus be seen as a call for action, inviting intelligence officers, perhaps for the first time, to join the ongoing debates around the future interpretation of the Geneva Conventions.

&!(#$!""#"J2J#%5",'-$b.<(+1"F=$;4"1#2"1"&'(%#A/1%=+"#)1."(()2"1A":#.4"#?&/9("+%#.4'.#

'&)%"#'1-#4/I# .4",# %/(L"# .4"+=$ \RO[U$ 4*$#''+,#*/#&!*2&L!$n'&:#/#$ [_S=$[__$ ORSP`U#O/,0,-F$'-$,-01"7,13$;1$/.-5*/015$3,0;$'$&1-,."$<'-5$J.31"$F1-1"'<$,-$0;1$6#:#$A,<,0'"8$3;.$&',5$'+.*0$0;1$3'"$,-$4"'cV$d4$$3'&$0"*&0,-F=$A'8+1$-'u71#$4-01<<,F1-/1$,-(."A15$A1$0;'0$4"'c$;'5$31'J.-&$.($A'&&$51&0"*/0,.-$'-5$0;18$31"1$+*,<5,-F$'$/'&1$0.$F.$0.$3'"#$4($4$;'5$+1001"$*-51"&0..5$g0;1h$,-01<<,F1-/1=$4$3.*<5$;'71$'&@15$0.*F;1"$c*1&0,.-&#$4$A'8$;'71$A'51$'$5,((1"1-/1#fU#$$

184 See Manjikian, supra note 134, at 45. 185 Id.