THOMSON REUTERS
LEGAL EXECUTIVE INSTITUTE
ARTIFICIAL INTELLIGENCE IN LAW:
THE STATE OF PLAY 2016
Artificial Intelligence (AI) and its impact and applications in the legal
profession is examined in this white paper by Michael Mills, Co-Founder &
Chief Strategy Officer of Neota Logic, a provider of intelligent software.
Mills analyzes AI – what the author calls a “big forest of academic and
commercial work around ‘the science and engineering of making intelligent
machines’” – and how AI is being implemented in legal areas such as
e-discovery, legal research, compliance, contract analysis, case prediction
and document automation.
by Michael Mills
2
ARTIFICIAL INTELLIGENCE IN LAW: THE STATE OF PLAY 2016
INTRODUCTION
Google Plays Go, Wins! No, that’s not another unpublished Dr. Seuss book. It’s the
dramatic outcome of artificial intelligence research at Google’s Deep Mind subsidiary,
whose AlphaGo program recently won five straight games against the top-ranked Go
master in Europe. The game of Go is 2,500 years old and, despite its simple rules, is many
orders of magnitude more complex than chess.
What is most remarkable about AlphaGo’s victory is that AlphaGo was not “taught” how
to play Go. Instead, its multilayer neural network learned how to play, and then how to
win, by playing millions of games and observing the winning strategies.
Ten years ago, IBM Deep Blue defeated the reigning world champion chess player. Five
years ago, IBM Watson defeated the two best Jeopardy players. One year ago, Google
Deep Mind learned to play, and win, 46 old Atari arcade games. Today, Deep Mind plays
Go, wins. (Facebook AI Research is playing Go too, and you can watch.)
These stunning and rapid advances in software that does what humans do, but better,
invite not only an optimistic question – what next? – but also a worried warning. In an
editorial accompanying publication of the AlphaGo research, the journal Nature wrote:
As the use of deep neural network systems spreads into everyday life – they are
already used to analyze and recommend financial transactions – it raises an
interesting concept for humans and their relationships with machines. The
machine becomes an oracle; its pronouncements have to be believed.
When a conventional computer tells an engineer to place a rivet or a weld in a
specific place on an aircraft wing, the engineer – if he or she wishes – can lift
the machine’s lid and examine the assumptions and calculations inside. That is
why the rest of us are happy to fly. Intuitive machines will need more than trust:
they will demand faith.
So, what does this mean for law?
The other day, a search for “artificial intelligence in law” produced 86,400 results from
just the News section of Google’s vast index. From the Web as a whole, 32.8 million
results, and from Videos – 261,000, beginning with Jude Law’s role as Gigolo Joe in the
movie “A.I.” (thank you, RankBrain).
The first News story was “Law firm bosses envision Watson-type computers replacing
young lawyers,” reporting on the answers to one question in the recent Altman & Weil
survey of law firm leaders. As wittily argued by Ryan McClead, “the question is flawed on
many levels [and] … it’s time to cut the hysteria surrounding artificial intelligence in law.”
Yes, there’s something going on here. But we need to parse the pile a bit. What is Artificial
Intelligence (AI)? What is AI doing in law? Who is doing it? And where is it headed?
WHAT IS THIS THING CALLED AI?
AI is a big forest of academic and commercial work around “the science and engineering
of making intelligent machines,” in the words of the person who coined the term artificial
intelligence, John McCarthy. A thorough and hype-free review of AI in business was
published recently by Deloitte, “Demystifying Artificial Intelligence,” suggesting the term
“cognitive technologies” to encourage focus on the specific, useful technologies that
emerge from the broad field of AI.
3
ARTIFICIAL INTELLIGENCE IN LAW: THE STATE OF PLAY 2016
However labeled, the field has many branches, with many significant connections and
commonalities among them. The most active today are shown here:
Planning
Robotics
Expert Systems
Supervised Machine Learning (ML)
Unsupervised
Deep Learning
Vision
Image Recognition
Machine Vision
Speech
Speech to Text
Text to Speech
Natural Language
Processing (NLP)
Content Extraction
Classification
Machine Translation
Question Answering
Text Generation
Artificial
Intelligence
(AI)
Lawyers do not need robots or machine vision, but other branches of AI are indeed useful.
Practical use of cognitive technologies in legal services is by no means new, nor did
it begin when IBM’s general counsel predicted that Watson could pass the bar exam
by 2016.
HARD AT WORK IN LAW
Artificial intelligence is hard at work in the law – for example, in legal research,
e-discovery, compliance, contract analysis, case prediction, and document automation –
though often there is no “AI Inside” label on the box.
Machine learning, expert systems, and other AI techniques enable lawyers to devote more
of their time to more valuable (and interesting) work. Mining documents in discovery and
due diligence, answering routine questions, sifting data to predict case outcomes, drafting
contracts – all are faster, better, cheaper, and becoming more so with the assistance of
intelligent software.
LEGAL RESEARCH
Lexis® and Westlaw® have applied natural language processing (NLP) techniques to legal
research for 10-plus years. No doubt Bloomberg BNA does as well. After all, the core NLP
algorithms were all published in academic journals long ago and are readily available.
The hard (very hard) work is practical implementation against good data at scale. Legal
research innovators like Fastcase and RavelLaw have done that hard work, and added
visualizations to improve the utility of results.
Recently, ROSS Intelligence has been applying IBM Watson’s Q&A technology to
legal research on bankruptcy topics, after winning a finalist spot in an IBM Cognitive
Computing Competition among 10 universities. After building and training the data set,
ROSS invites users to evaluate search results and feeds those evaluations back to
4
ARTIFICIAL INTELLIGENCE IN LAW: THE STATE OF PLAY 2016
the engine to continue tuning (the essence of machine learning) in the manner of
recommendation engines at Netflix and Amazon® as well as Google® feedback loops,
based on what we do with the search results we’re shown. No date for commercial
release of ROSS has been announced.
Last October, Thomson Reuters, publishers of Westlaw (and incidentally, the Legal
Executive Institute blog), announced a collaboration to use Watson across Thomson
Reuters information businesses. Although nothing was said publicly about Thomson
Reuters specific plans for Watson, one could speculate that the vast trove of legal
content in Westlaw and the army of subject-matter experts in the company could
together do impressive things to improve legal research. Watson needs big data and
training, at least initially, by people: Thomson Reuters has both.
On February 1, at a private “innovation summit,” Thomson Reuters teased the legal
industry with hints that Watson Esq. will ride into town with a beta service for financial
services regulation by the end of this year. Jean O’Grady’s commentary is, as usual,
acute.
Take note of the timeline: Even a company with the immense resources of content and
expertise of Thomson Reuters, even in partnership with IBM, needs more than a year to
get to beta with an AI legal research product. Why? Because neither AI nor Watson is
magic. It takes time, human expertise, and painstaking effort to assemble useful data
sets, analyze the content, train the algorithms, and test the results. The broader the
targeted topic, the greater the effort. For perspective, the IBM Jeopardy team’s account
of their work is excellent.
ELECTRONIC DISCOVERY
Technology-assisted review (TAR, or predictive coding) uses natural language and
machine learning techniques against the gigantic data sets of e-discovery. Recommind,
Equivio (now part of Microsoft®), Content Analyst, and other vendors develop or license
these tools. TAR has been proven to be faster, better, cheaper, and much more consistent
than human-powered review (let’s have another initialism: HPR). See, for example, the
paper by University of Waterloo’s Gordon V. Cormack and Wachtell, Lipton, Rosen &
Katz’s Maura R. Grossman, Evaluation of Machine-Learning Protocols for Technology-
Assisted Review in Electronic Discovery. (The story of Grossman and Cormack’s work was
well told recently by Susan Beck in The American Lawyer.)
Yes, it is assisted review, in two senses. First, the technology needs to be assisted; it
needs to be trained by senior lawyers very knowledgeable about the case. Second, the
lawyers are assisted by the technology, and the careful statistical thinking that must be
done to use it wisely. Thus, lawyers are not replaced, though they will be fewer in number.
Done right, TAR is both powerful and reliable. Doing it right isn’t easy. One needs to
understand the principles, and even some of the statistical mathematics, especially when
appearing in court to argue that the outcomes are defensible and consistent with the
standards of Federal Rule of Civil Procedure 26 and comparable rules in other courts.
A good place to start the journey to understanding is TAR for Smart People, a book by
John Tredennick, one of the pioneers of e-discovery (and legal technology generally).
TAR for Smart People is a superb guide to a critical and often misunderstood topic.
The book is clear but technically deep, founded on fact, balanced, and engaging.
Who knew statistical sampling could be fun?
In scale and impact on costs, TAR is the success story of machine learning in the law. It
would be even bigger but for the slow pace of adoption by both lawyers and their clients.
5
ARTIFICIAL INTELLIGENCE IN LAW: THE STATE OF PLAY 2016
OUTCOME PREDICTION
Lex Machina, after building a large and fine-grained set of intellectual property (IP) case
data, uses data mining and predictive analytics techniques to forecast outcomes of IP
litigation. Recently, it has extended the range of data it is mining to include court dockets,
enabling new forms of insight and prediction. For example, the Motion Kickstarter
enables:
attorneys [to] view granted motions with denied motions to see what’s working
and what’s not. Enter a judge’s name and motion type and instantly view the
judge’s recent orders on that motion type, as well as the briefing that led up to
those orders.
LexPredict has built models to predict the outcome of Supreme Court cases, at accuracy
levels challenging experienced Supreme Court practitioners. Premonition says they are
using data mining and other AI techniques “to expose, for the first time ever, which
lawyers win the most before which judge.”
Perhaps Huron’s Sky Analytics and the new AIG spinoff, Legal Operations Company, can
use their big databases of law firm case and billing data to offer outcome predictions as
well as cost and rate benchmarks.
SELF-SERVICE COMPLIANCE
Neota Logic applies its hybrid reasoning platform, which combines expert systems and
other artificial intelligence techniques, including on-demand natural language processing
(NLP) and machine learning, to provide fact- and context-specific answers to legal,
compliance, and policy questions. (Disclosure: I am co-founder and chief strategy officer
of Neota Logic.)
ComplianceHR, a joint venture of Littler Mendelson and Neota Logic, offers a suite of
Navigator applications to assist human resources professionals in evaluating independent
contractor status, overtime exemption, and other employment law issues. Foley & Lardner
uses expert systems technology to power its Global Risk Solutions service, an “integrated
Foreign Corrupt Practices Act (FCPA) compliance solution that addresses each of the
‘hallmarks’ of an effective anti-corruption compliance program identified” by regulatory
authorities.
CONTRACT ANALYSIS
General counsels recognize that their high priorities of risk management and cost
reduction are served by understanding and managing the rights, obligations, and risks
in a company’s contracts, and rationalizing the processes by which contracts are initiated,
negotiated, drafted, and managed through their life cycle, from execution to expiration.
Natural language processing, machine learning, and other AI techniques are being
applied to many aspects of the contract life cycle, including discovery, analysis, and
due diligence.
For example:
• Kira Systems reports that contract review times in due diligence can be reduced by
20 to 60%.
• KM Standards can “identify common clauses, agreement structure, standard clause
language, and common clause alternatives” in a set of contracts.
ARTIFICIAL INTELLIGENCE IN LAW: THE STATE OF PLAY 2016
• RAVN’s cognitive computing platform, the Applied Cognitive Engine (ACE to its friends),
will “read, interpret, and summarize” key information from contracts.
• Seal Software can crawl a network to discover, and then classify, all of a company’s
existing contracts.
Contract analytics is well on its way to being a success story for machine learning in
the law.
IS IT TIME TO GET IN THE GAME?
Many, perhaps most, law firms choose not to be early adopters of new technologies. Likely,
that is not because they have read about the rewards of being a “fast follower” instead of
a “first mover.” Rather, they are lawyers – educated to precedent, alert to their peers, wary
of failure, and hence, reluctant to experiment.
However, as I hope this quick tour has shown, notwithstanding the chatter and excitement
about the arrival of Watson in Law Land, the techniques of cognitive technologies are
robustly at work in the trenches of law practice, doing useful work today – improving
service to clients, reducing costs, and creating new opportunities for firms.
THE FUTURE?
More, and better, of course. Cognitive technologies in the law are riding a wave of eversmarter
algorithms, infinite scaling of computer power by faster chips and cloud-clustered
servers, intense focus by companies led by seasoned experts, and an ever-greater demand
from clients for cheaper, faster, better services.
Note that cheaper is only one of the three words. Faster is important – companies
measure cycle time, time to market, and other indicia of speed throughout their
businesses, and increasingly expect their lawyers to do the same. And better is critical –
big companies face ever-growing regulatory and operational complexity, for which
traditional legal services on the medieval master craftsman model are simply inadequate.
To meet those needs, only technology-enabled services will do the job. And artificial
intelligence is driving those changes.
© 2016 Thomson Reuters S031401/3-16
Michael Mills
Michael Mills is the co-founder & chief
strategy officer of Neota Logic, served as chief
knowledge officer of Davis Polk & Wardwell,
and was a partner at Mayer Brown.
ABOUT THOMSON REUTERS LEGAL EXECUTIVE INSTITUTE
The Legal Executive Institute brings together people from across the
legal industry to ignite conversation and debate, make sense of the latest
events and trends, and provide guidance as you confront the opportunities
and challenges that these changes present. Visit us at: .
