Download - Machine Intelligence and the Legal Profession - John O. McGinnis - June 2016 OECD discussion
PRESENTATION BY JOHN O. MCGINNISGEORGE C. DIX PROFESSOR IN CONSTITUTIONAL LAW AT NORTHWESTERN PRITZKER SCHOOL OF
LAW
Machine Intelligence and the Legal Profession
Accelerating Computational Power
Machine Intelligence has continued to become ever more powerful
Moore’s Law has held for 40 years
Improvements in software and connectivity are force multipliers
Accelerating Computational Power
Exponential growth is likely to continue through other means, like carbon nanotubes and optimal computing
Accelerating Computational Power
Machine intelligence has moved from formal systems to systems with more uses in everyday life and in the profession of law1997 – Big Blue beats World Chess Champion Gary Kasparov2011 – Watson beats Jeopardy Champion Ken Jennings2013 – IBM builds a division around Watson’s analytics2016 – Law firms hire Ross, a service based on Watson’s analytics
Predictive Coding
Source: Lit i View
After being trained by lawyers on a sample of relevant documents, machines are responsible for finding other documents in discovery
Predictive coding is advancing One 2011 study indicated that:
Manual reviewers identified between 25% and 80% of relevant documents, while technology-assisted review returned between 67% and 86%
Technology-assisted review required human review of just 1.9% of the documents
Courts are increasingly granting requests from parties to use predictive coding in litigation
Common methodologies: Concept Searching Contextual Searching Metadata Searching
Predictive Coding
Law firms have problems creating disruptive technology in their field
It is not lawyers, but other businesses that are leaders
Important Players:
Predictive Coding
Legal Search
Computerized legal search has been around for 40 years, but it has been focused on key words
Now the movement is toward semantic search Using IBM’s Watson, a team of students at University
of Toronto launched ROSS Intelligence in 2015 with support from global law firm Dentons
LexisNexis TotalPatent includes a semantic search option
Moneyball comes to lawDefined by the Gartner IT Glossary as:
“Any approach to data mining with four attributes: An emphasis on prediction; Rapid analysis measured in hours or days (rather than the
stereotypical months of traditional data mining); An emphasis on the business relevance of the resulting
insights (no ivory tower analyses); and An increasing emphasis on ease of use, thus making the
tools accessible to business users.”Law firms can use predictive analysis for:
Targeting lucrative clientele Pricing matters based on past performance Predicting outcomes of litigation
Legal Predictive Analytics
Legal Predictive Analytics
In 2015 LexNexis bought out Lex Machina, a company with substantial data
Legal Predictive Analytics
Docket Alarm’s Patent Trial and Appeals Board analytics program allows users to generate reports on judges, as well as parties, firms, and technology areas to see who settles the most and at what stage.
Transactional Documents
Legal Document generation is occurring at both high and low ends
Legalzoom creates documents, such as wills, for the masses
Startup Documents has created a program to generate and store documents for startups wishing to incorporate, grow, and maintain their businesses
RocketLawyer provides individuals and businesses with legal document generation and enforcement services
Linear v. Accelerating World
Once computation gets into a space, it does not stopImproves search, document generation, predictionEven brief writing may be invaded in the decades to
come
Evidence for Effect Already
1. Decline of incomes of “small” lawyers2. Stagnant associate salaries3. Drop in talented people applying to law
schools – their own future prediction4. Huge increase in start-ups in legal space
The Stanford Center for Legal Informatics hosts a curated list of 551 companies “changing the way legal is done”
The list includes 153 document automation companies, 42 legal research companies, and 38 analytics companies
Evidence for Effect Already
Long-term Changing Composition
Technology is more likely to displace lawyers in low-value and simple matters Innovation happens at the low-end first i.e., drafting standard legal forms, areas of law where
patterns and past data can predict the outcome of a case Lawyers in relatively stable areas of law like trusts and
estates will be severely disrupted
Technology is not likely to displace lawyers in high-value and complex matters Specialists in novel areas of law will still be needed Lawyers in areas of fast-changing law, like much of financial
regulation, will be largely unaffected
Machines cannot substitute for trial lawyers – but predictive analytics will help convergence on value of lawsuits, reducing the need for trials
Premium on psychology and bonding, getting clients to take sensible actions
Long-term Changing Composition
Embracing Technology
Instead of trying to slow down computation, lawyers should relax ethical rules to permit others to own law firms. See changes in Britain and Australia.
Only firms with substantial capital will be able to innovate
Roadblocks to Innovations
1. Unauthorized practice of law rules not likely to slow innovation
Legislators have pushed back on behalf of constituents In any event, most computerized service can be used as inputs into
lawyers’ work and the effect will mostly be the same Use of technology can also be placed offshore to evade restrictions
2. Rules structuring legal partnerships pose great threat to innovation
Model Rules of Professional Conduct 5.3, 5.4, 5.5, and 5.7 restrict the ability of companies to earn profits from providing legal services
Makes it difficult for law firms to raise capital to innovate and collaborate with innovators
Makes it more likely that lawyers will try to slow down innovation