1. Overview of ICT’s impact on employmentICT progress has 2 opposite effects: an employment
replacement effect, in which ICT replaces operations previously done by humans, and an employment cre-ation effect, in which increased added value and the cre-ation of new business domains, brought on by ICT ap-plication and use, will increase employment (Figure 4-1-1-1).
The displacement of employment by ICT tends to be
viewed negatively. But it can be viewed positively as well, since the application and use of ICT can improve labor productivity at enterprises and create considerable add-ed values. And in Japan, where the declining birthrate and aging population is forecast to reduce the working population, the application and use of ICT will likely be effective in addressing labor supply limitations.
2. Changes in the forms of employment replaced by ICTPrior research on employment replacement by ICT
has clarified that ICT progress to date has replaced hu-mans in routine types of work operations (for example, accounting operations and production processes) but not in non-routine types of work operations (for exam-ple, research positions and sales positions) or manual
operations (for example, sales promotion operations). There is growing consensus, however, that the rapid ad-vance of AI technology and robotics in recent years will, in the future, replace humans with machines in non-rou-tine types of knowledge operations and complex manual operations too.
3. ICT and new work styles ICT progress is bringing enormous changes to how
people work. For example, telework is making it possi-ble for people to work independent of location. And the proliferation of the sharing economy and digital fabrica-
tion are creating more opportunities for individuals to work on demand without belonging to an organization. Such new work styles will allow people to choose more diverse and flexible working arrangements, which is ex-
Section 1 ICT Progress and Employment
Much recent research around the world often speaks of artificial intelligence (AI) in negative terms, as something with the potential to replace human workers. In this chapter, we present the findings of surveys on this issue of AI and employment given to workers in the United States and Japan and on the sentiments of experts in Japan. We also map out the evolution of AI and suggest measures that will be required in the future to deal with AI.
Prepared from the “Study Report on Issues and Solutions to Regional Employment Creation with ICT” (2015), MIC
Figure 4-1-1-1 Impact of ICT on employment
Advances in ICT
Higher productivity at enterprises
Cut employment Add employment
(Impacts on employment)
Creation of new business domains
Higher added valueLower costs
・ICT will tend to replace jobs with routine types of work operations.
・As ICT helps enterprises grow and expand, they will add more jobs.
・ICT creates new business domains, which will create new jobs.
40
Part 1
Chapter 4ICT Progress and Future Work Styles
CW6_AY057D04.indd 40 2016/12/08 10:25:25
pected to improve their work-life balance.
(1) Teleworka. State of telework adoption by Japanese enterprises
According to the MIC’s 2015 Communications Usage Trend Survey, only 16.2 percent of enterprises answered they had “implemented” telework5 in Japan, a rate that increased to just around 20 percent when adding in en-terprises that answered “have not implemented but have specific plans to do so.”
b. Japan-U.S. comparison of worker perceptions of teleworkWhat do workers think of telework? We surveyed
worker monitors in Japan and the United States about their intentions to make use of telework.
The gap in telework intentions between the 2 coun-tries was significant, with over 60 percent of U.S. work-ers answering “I would like to try telework,” versus 30 percent of Japanese workers. The intention to make use of telework in Japan was only half that in the United States (Figure 4-1-3-1).
(2) Sharing economyConsumers in many countries are starting to accept
new forms of sharing economy services powered by ICT.
We surveyed worker monitors in Japan and the Unit-ed States about their intentions to work in sharing-econ-omy style services (i.e., sharing economy work styles). Over 40 percent of workers in the United States an-swered “I would like to try a sharing economy work
style.” This was more than twice the rate in Japan, where just over 20 percent of workers gave the same answer (Figure 4-1-3-2).
(3) Digital fabricationDigital fabrication technology produces physical ob-
jects based on digital data. 3D scanners, 3D CAD sys-tems, and other measurement machines convert peo-ple’s ideas or a person’s physical dimensions into digital data. The digital data are loaded into 3D printers, laser cutters, and other digital fabrication machines to pro-duce the final object.
We surveyed worker monitors in Japan and the United States about their intentions to make use of digital fabrica-tion. More than half of U.S. workers answered “I would like to try” digital fabrication, which was more than dou-ble the response rate in Japan, where around 20 percent of workers gave the same answer (Figure 4-1-3-3).
(4) New work styles (summary)ICT advancements are set to enable many new work
styles, such as telework, the sharing economy, and digi-tal fabrication. We asked workers in Japan and the Unit-ed States which of these new work styles they found at-tractive. For all work styles, more U.S. workers found them attractive than Japanese workers. This suggests Japanese workers are not looking for change in their work styles. This reluctance is likely one factor for the low intention to take advantage of telework, the sharing economy, and digital fabrication (Figure 4-1-3-4).
5 Telework includes working from home, working at satellite offices, and mobile work.
Prepared from the “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-1-3-1 Workers’ intention to make use of telework (Japan / United States)0 20 40 60 80 100 (%)
US(n=1,105)
Japan(n=1,106)
I would like to try I would not like to try
30.1
63.3
69.9
36.7
Prepared from the “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-1-3-2 Workers’ intention to make use of sharing economy work styles (Japan / United States)
21.5
44.3
78.5
55.7
0 20 40 60 80 100
Japan(n=1,106)
US(n=1,105)
(%)
I would like to try I would not like to try
41
Part 1
CW6_AY057D04.indd 41 2016/12/08 10:25:28
1. Progress in artificial intelligence research(1) What is artificial intelligence?
AlphaGo shocked the world when it defeated the 9-dan rank professional Go player, Lee Sedol, 4 games to 1 in a five-match competition in March 2016. Compared to chess or shogi, Go has a bigger board and infinitely more board combinations. So many believed it would take years before artificial intelligence (AI) could sur-pass human abilities in Go, but AlphaGo’s victory showed the world that AI progress had reached another level.
a. What really is artificial intelligence?The term AI is quite well recognized now, but the
world first learned of it relatively recently at the Dart-mouth Conferences in 1956. In broad terms, AI can be described as “the science and engineering of making intelligent machines, especially intelligent computer programs,” but researchers are divided on the definition of AI. The reason for this, it has been pointed out, is that there is no proper definition of “intelligence” to begin with, so defining “artificial” intelligence is rather impos-
sible. For our purposes, we have not chosen a particular
definition for AI and simply describe AI as “the science and engineering of making intelligent machines, espe-cially intelligent computer programs.”
(2) History of artificial intelligence researchAI research has been ongoing since the 1950s, under-
going several booms and winters in the process. AI is now in the limelight again, in its third boom.
a. Looking back on the first 2 artificial intelligence boomsIn the previous 2 AI booms, society’s expectations for
AI exceeded the technical limits AI could achieve at the time, and the booms fizzled out when this discrepancy became clear. Given this history, it has been said, regard-ing the present third boom, that people need to recognize there will be a gap in the latent potential that AI technical development and implementation, at its most successful, can achieve and what is forecast to be realistic.
Section 2 Present and Future of Artificial Intelligence (AI)
Prepared from the “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-1-3-3 Workers’ intention to make use of digital fabrication (Japan / United States)
20.8
53.9
79.2
46.1
0 20 40 60 80 100
Japan(n=1,106)
US(n=1,105)
(%)
I would like to try I would not like to try
Prepared from the “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-1-3-4 Work styles viewed as attractive (Japan / United States)
25.425.4
20.020.0
11.511.5
13.713.7
7.37.3
4.24.2
42.342.3
36.836.8
5.45.4
10.210.2
0.00.0
30.230.2
49.049.0
20.320.3
20.920.9
14.814.8
18.818.8
11.711.7
57.857.8
54.054.0
16.316.3
20.020.0
0.90.9
8.88.8
0 10 20 30 40 50 60 (%)
Work while having a child, raising children,or caring for parents
Work at more than one job
Work together with people from differentcountries and with different languages
Lend out normally unusedthings or vacant locations
Work using connections andexchanges on social media
Work using reviews and opinionson social media or review sites
Work during the times you like
Work in the locations you like
Work that focuses on creating or shapingconcepts and ideas while shortening the
time spent on design and prototyping
Easily make your concepts and ideas take shapewithout needing to belong to an organization
Other
None of the work styles above are attractive
Japan(n=1,106) US(n=1,105)
42
Part 1
CW6_AY057D04.indd 42 2016/12/08 10:25:31
2. Expanding artificial intelligence applications(1) Examples of using and applying artificial intelligence
Expectations for AI, particularly deep learning, are that application fields will grow, as identification and forecasting accuracies improve, and that the integration of multiple technologies will fulfill functions people have long wanted. AI will certainly develop along multiple paths, but the table below lists the assumptions of ex-perts in the field (Figure 4-2-2-1).
At the same time, it is up to humans to decide how AI will be used, and in which fields, as well as how AI will not be used. Efforts are underway to form a consensus on preventing the use of AI in ways that will harm soci-
ety — ELSI (ethical, legal, and social issues).
(2) Applications of artificial intelligencea. Promising fields for artificial intelligence applications
We surveyed experts in the field on promising fields for AI applications. The results indicated there are quite strong needs for AI applications in fields where social issues are expected to be solved. These include more advanced health examinations, autonomous operation of public transportation, route selection for emergency ve-hicles and ambulances, and easing of traffic congestion and crowding.
1. Artificial intelligence and employment(1) State of artificial intelligence adoption in workplaces
We asked workers in Japan and the United States about the state of AI adoption in their present workplac-es. The results indicated AI workplace adoption at the present time is more advanced in the United States than in Japan (Figure 4-3-1-1).
Next, we asked workers who answered “AI has al-ready been adopted in my workplace” or “AI is planned to be adopted in my workplace” what role or function AI plays (or will play) in the workplace. The most common answer among Japanese workers was “role or function is to reduce current labor levels,” at 41.0 percent, followed
Section 3 Influence of AI Progress on Employment, etc.
Prepared from the “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-3-1-1 State of current and planned AI workplace adoption
5.3
3.1
8.3
5.6
16.5
63.2
51.9
26.3
17.9
0 20 40 60 80 100 (%)
AI already adopted and I have used itAI already adopted, but I haven’t used it yetAI not adopted at the present time, but it is planned to be adopted (planning or under consideration)AI not adopted at the present time and no adoption plansDon’t know
Japan(n=1,106)
US(n=1,105)
1.9
(Source) “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-2-2-1 Development of AI and evolution of AI applications
・Autonomous driving・Distribution (last mile)・Robotics
・Big data・Crime prevention / surveillance
・Advertising・Medical diagnoses from images2014
2015
・Education・Secretarial work・Assistance with white-collar jobs
・Translation・EC for overseas markets
・Entry into society・Housework and caregiving・Substitute for emotional labor
2020
Behavior and planning
Multimodal abstraction
Image recognition
Accumulation of even more knowledge
Association with language
Abstraction based on behavior
Autonomous action plans
Recognition of emotionsBehavior predictionsSituational awareness
More accurate recognition
Large-scale knowledge
understanding
Language understanding
Huge improvement in situational
awareness ability
Impact on societyTechnological advance Improved technologiesYear
43
Part 1
CW6_AY057D04.indd 43 2016/12/08 10:25:34
by “role or function is to supplement labor shortages” and “role or function is to raise efficiency and productiv-ity of existing operations,” both at 35.0 percent (Figure 4-3-1-2).
(2) Perceptions on artificial intelligence adoptionWe asked workers in Japan and the United States
about the possibility of AI becoming their work partner (either as supervisor, colleague, or subordinate) and how much they oppose such a possibility. In the United States, many workers “strongly oppose” or “somewhat oppose” AI becoming their colleague or subordinate, whereas many Japanese workers opposed AI becoming their supervisor (Figure 4-3-1-3).
(3) Envisioned impact of artificial intelligence adoption on em-ploymentThere is broad public consensus that AI will engender
2 effects — higher operational efficiencies and produc-tivity and the creation of new operations and businesses — and that these 2 effects will transform the volume of tasks that constitute employment.
The AI effect of higher operational efficiencies and productivity will reduce the amount of tasks in jobs where AI is adopted. On the other hand, the AI effect of creating new operations and businesses is expected to
increase the amount of tasks in newly created jobs. Spe-cifically, AI will create jobs needed to introduce and ex-pand AI and new jobs that make use of AI. The ideal for society is that the newly created task amounts exceed the reduced task amounts, resulting in a net gain in task volumes. Therefore, the creation of new operations and businesses through AI has a significant and large role to play.
There are 4 envisioned impacts on employment that will be brought about by such task volume transforma-tions: (1) partial substitution of employment; (2) comple-mentation of employment; (3) maintenance and expan-sion of employment by directly linking to industrial competitiveness; and (4) improvement of the working environment of women and the elderly, etc. (Figure 4-3-1-4).
a. Impact of growth in artificial intelligence on employment in JapanWe asked experts in the field about what impact AI
adoption and growth will have on employment in Japan. Twenty-three of 27 experts said “AI can make up short-ages in the labor supply caused by the continuation of low birthrates and aging population” (Figure 4-3-1-5).
b. Impact of artificial intelligence workplace adoption on operationsWe asked workers in Japan and the United States
Prepared from the “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-3-1-2 Role or function of AI
35.0
41.0
35.0
31.6
26.5
6.8
6.8
0.0
19.2
39.0
48.6
30.0
19.5
13.5
11.1
2.1
0 5 10 15 20 25 30 35 40 45 50 (%)
Supplement labor shortages
Reduce current labor levels
Raise efficiency and productivity of existing operations
Raise value provided by existing operations(quality, consumer satisfaction, etc.)
Create operations with new value that didn’t exist before
Raise motivation and satisfaction in carryingout existing operations
Raise motivation and satisfaction in carrying out new operations
Other
Japan(n=117) US(n=333)
Prepared from the “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-3-1-3 Possibility and opposition to AI as a work partner
Japan(n=1,106)
US(n=1,105)
Supervisor Colleague Subordinate
27.5%
46.4%
20.7%
5.3%
Strongly oppose Somewhat oppose Do not really oppose Do not oppose at all
33.9%
29.1%
21.1%
15.9%
25.2%
37.3%
28.3%
9.1%
19.9%
28.3%34.0%
17.8% 17.3%
25.0%
35.0%
22.7%
28.7%
41.0%
22.9%
7.4%
44
Part 1
CW6_AY057D04.indd 44 2016/12/08 10:25:34
about the impact AI adoption in their current workplace has had on operations. More U.S. workers answered “has had a very large impact” or “has had some impact” than Japanese workers (Figure 4-3-1-6).
We next asked about the impact AI workplace adop-tion has had on the range of operations. More U.S. work-ers answered that operations had increased — total of “increased significantly,” “increased somewhat,” and “increased slightly” — than answered that operations had decreased — total of “decreased significantly,” “de-
creased somewhat,” and “decreased slightly.” Thus, U.S. workers are inclined to view AI adoption as expanding operations. The opposite tendency was seen among Jap-anese workers, with those answering that operations had decreased exceeding those answering that opera-tions had increased by 5.9 percentage points. Thus, com-pared to their U.S. counterparts, Japanese workers are inclined to view AI adoption as narrowing operations (Figure 4-3-1-7).
(Source) “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-3-1-4 Envisioned impact of AI adoption on employmentIntroduction and expanded use and application of artificial intelligence (AI)
Influ
ence
on
empl
oym
ent
Complementation of employment
Maintenance and expansion of employment by drectly
linking to industrial competitivenessImprovement of the working
environment of womenand the elderly, etc.
Chan
ges
in th
e ta
sk v
olum
e
Partial substitution of employment
1 2 3 4
Occupations with low potential for
mechanization
Occupations with low potential for
mechanization
Occupations with high potential for mechanization
Introduction of AIExpanded use and
application of AI
Due to the effect of AI on improvingoperational efficiency and productivity,
the task volume of occupations with high potential for mechanization decreases.
Due to the effect of AI on creating new businesses, the task volume of newly
created occupations increases.
Newly created occupations
The task volume increases due to the following two kinds of work:“work required for introducing and penetrating AI” and
“new work utilizing AI.”
AI will not take over all work, in other words, it will not deprive workers of their employment. Instead, it will only take over some tasks for which use of ordinary workers would be costlier for achieving the same level of productivity.
The labor supply shortage caused by progress of the declining birth rate and aging population will be complemented by AI, people working with AI, and people whose task volume has decreased due to AI.
Employment will be maintained and expanded as a result that companies which were early in using and applying AI increase their industrial competitiveness. (However, for Japanese companies, a delay in response to digitization or operational process optimization tends to become an obstacle to the introduction, use and application of AI.
By shifting to highly productive work that efficiently uses AI, people will be able to work flexibly without always being required to work long hours, which will expand working opportunities for women and the elderly, etc.
Prepared from the “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-3-1-5 Impact of AI adoption and penetration on employment in Japan
23
18
16
9
5
4
3
8
0
0 5 10 15 20 25AI can make up shortages in the labor supply caused by the continuation of low birthrates and aging population
Raise operational efficiencies and productivity, leading to shorter working hours
Create new markets and expand employment opportunitiesForm agreeable employment conditions for women,
non-Japanese people, older people, and others
Many people will lose employment and the unemployment rate will increase
Other
No particular impact on employment (n=27)
(人)
Dependency on AI will grow and people will be less creative in their work
AI will leave many people behind and create employment mismatches
Prepared from the “Study Report on the Impact of ICT progress on Employment and Work Styles” (2016), MIC
Figure 4-3-1-6 Impact of AI workplace adoption on operations
12.6
30.8
33.5
39.8
31.6
17.0
22.3
12.4
0 2010 30 50 7040 60 80 90 100 (%)
Has had a very large impact Has had some impactHasn’t had much impact Has had no impact
Japan(n=1,106)
US(n=1,105)
45
Part 1
CW6_AY057D04.indd 45 2016/12/08 10:25:35
1. Facing further artificial intelligence expansion and growth(1) Supporting artificial intelligence
What do workers in Japan and the United States think of preparing for and dealing with future AI growth?
Many U.S. workers said “I’m preparing for AI from the standpoint of using AI so I can continue my current job and duties, such as acquiring AI knowledge and skills.” This suggests an attitude that puts importance on pre-
paring for and dealing with AI growth so that workers can master AI and apply it to their current jobs and du-ties. Conversely, in Japan, more than half of workers said “I’m not doing anything specific to prepare for or deal with AI.” This attitude prompts concerns that people will be left behind as AI expands and grows (Figure 4-4-1-1).
2. Future education and human resources development(1) Needs in study environments and support structures
In both Japan and the United States, there were strong needs for “enriched education and research at institutes of higher learning” and for “support structures for self-development at enterprises” in the area of study environ-ments and support structures, so that workers can ac-quire the skills they need to make use of AI (Figure 4-4-2-1).
At the same time, support for individual self-develop-ment alone is, naturally, insufficient to establish the con-ditions for AI adoption and usage. For example, Recruit Holdings has successfully laid the groundwork for all employees to make use of AI in 3 stages: (1) hiring AI
experts who take work from all workplaces across the organization; (2) increasing the number of AI experts and assigning AI experts to workplaces who work along-side people in the workplace; and (3) creating conditions for all employees to use AI through systemization.
As this example shows, when it comes to acquiring the necessary attributes and skills to use AI, it is more important that the enterprise lower the barriers to AI usage, while employees study to use AI in the work-place, than to establish usage conditions where people can learn AI skills through experience in the workplace (Figure 4-4-2-2).
Section 4 Changes in the Required Skills and the Types of Human Resources and Education Sought
Prepared from the “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-3-1-7 Impact of AI workplace adoption on range of operations
6.9%
11.4%
8.8%
39.9%
20.5%
8.5%
4.0%
11.9%
15.3%
11.9%31.4%
16.0%
8.8%
4.8%
Japan(n=859) US(n=968)
Range of operations have decreased significantlyRange of operations have decreased somewhatRange of operations have decreased slightlyRange of operations have not changedRange of operations have increased slightlyRange of operations have increased somewhatRange of operations have increased significantly
Prepared from the “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-4-1-1 Preparations to handle future AI growth
13.1 11.4
28.0
7.7
51.2
26.922.9
46.7
18.122.8
0
10
20
30
40
50
60(%)
Japan(n=1,106) US(n=1,105)
I will
not
mak
e an
y pa
rticu
lar
resp
onse
or p
rep
arat
ions
.
Resp
onse
and
/ o
r pre
para
tions
invo
lvin
g b
eing
tran
sfer
red
or c
hang
ing
jobs
to a
noth
er li
ne o
f wor
k in
whi
ch I
can
utili
ze th
e kn
owle
dge
and
skill
s w
hich
I ha
ve d
evel
oped
.
Resp
onse
and
/ o
r pre
para
tions
invo
lvin
g b
eing
tran
sfer
red
or c
hang
ing
jobs
to a
noth
er li
ne o
f wor
k un
rela
ted
to th
e k
now
ledg
e an
d sk
ills
whi
ch I
have
dev
elop
ed.
Resp
onse
and
/ o
r pre
para
tions
invo
lvin
g c
ontin
uing
my
curr
ent w
ork
with
the
pos
ition
of u
sing
AI (a
rtific
ial i
ntel
ligen
ce)
thro
ugh
actio
ns s
uch
as le
arni
ng k
now
ledg
e an
d sk
ills
of A
I
Resp
onse
and
/ o
r pre
para
tions
invo
lvin
g b
eing
tran
sfer
red
or c
hang
ing
jobs
to a
noth
er li
ne o
f wor
k w
ith th
e po
sitio
n o
f usin
g AI
(arti
ficia
l int
ellig
ence
) thr
ough
act
ions
suc
h as
lear
ning
kno
w le
dge
and
ski
lls o
f AI
46
Part 1
CW6_AY057D04.indd 46 2016/12/08 10:25:36
3. Roles expected of government in spreading artificial intelligence We asked experts about the roles expected of govern-
ment in promoting AI research and development and spreading AI in society. Partly because the survey in-cluded many experts in the ICT field, including AI, a
majority of the experts indicated that the government should institute policies to promote practical AI applica-tions and the deployment of those applications (Figure 4-4-3-1).
(Source) “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-4-3-1 Role expected of government in promoting AI research and development and spreading AI in society
15 0 5 7
(n=27) (Unit: Experts)
Government should institute policies to promote practical AI applications and the deployment of those applications
Government should select fields in which to promote practical AI applications and the deployment of those applications and fields where the stability of employment and other issues are taken into account
Government should institute policies that prioritize the promotion of practical AI applications and the deployment of those applications while accounting for the stability of employment and other issues
Government should be neutral and leave the practical AI applications and the deployment of those applications to private-sector technological developments and the judgment of markets
(Source) “Study Report on the Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-4-2-2 Example of an enterprise setting up conditions for AI usage
Time
Work requestsWork requests
AI expertsAI experts AI expertsAI experts AI expertsAI experts
Provide AI support throughoutthe organization
Organize the AI support
Hire AI experts who take workfrom all workplaces across theorganization
Co-workingCo-working
AssistanceAssistance
Systemize the AI support
While increasing the number of AI experts, assign AI experts to workplaces who work alongside people in the workplace
Through systemization, thecompany creates conditions forall employees to use AI
AssistanceAssistance
AI system
No.
of p
eopl
e us
ing
AI in
the
com
pany
Dep
t. A
Dep
t. B
Dep
t. C
Dep
t. D
Dep
t. E
Dep
t. A
Dep
t. B
Dep
t. C
Dep
t. D
Dep
t. E
Dep
t. A
Dep
t. B
Dep
t. C
Dep
t. D
Dep
t. E
Prepared from the “Study Report on Impact of ICT Progress on Employment and Work Styles” (2016), MIC
Figure 4-4-2-1 Needs for a learning environment and support systems for learning skills for using AI
27.9
36.5
28.4 27.2
18.5 19.7
11.0
0.0
6.8
41.6
50.246.7
38.5
23.118.7
12.4
0.4 1.90
10
20
30
40
50
60(%)
Japan(n=680) US(n=937)En
richm
ent o
f edu
catio
n an
d re
sear
ch in
hig
her e
duca
tion
faci
litie
s
A su
ppor
t sys
tem
for s
elf-e
duca
tion
in c
ompa
nies
Educ
atio
nal a
nd tr
aini
ng p
rogr
ams
pro
vide
d by
an
outs
ide
serv
ice
Syst
emat
ic a
nd p
ract
ical
pro
mot
ion
of O
JT in
com
pani
es
Crea
tion
of n
atio
nal q
ualifi
catio
ns a
nda
natio
nal t
estin
g sy
stem
Hold
ing
of c
onte
sts
in v
olvi
ng c
ompe
titio
n o
f ide
as a
nd c
reat
ivity
Oth
er
A pa
rticu
lar l
earn
ing
envi
ronm
ent o
r s
uppo
rt sy
stem
is n
ot n
eces
sary
.
Lear
ning
inst
ruct
ion
at th
e m
anda
tory
educ
atio
n st
age
47
Part 1
CW6_AY057D04.indd 47 2016/12/08 10:25:36