For the last few years lawyers have heard a great deal about artificial intelligence (AI) and its effect on the practice of law. Although there is no cause for breathless claims of robot lawyers coming to take our jobs, the last few years have seen interesting new software applications to augment traditional, human lawyering.
The first wave of AI applications for law is beginning to gain maturity in the market. Perhaps even more importantly, law firms are starting to create their own data analysis tools using AI, the beginning of a “maker movement” in law.
Creating a New Tier of Legal Services
This new age of law firm innovation can create changes in the nature of legal services, moving from anecdotal to data-driven insights, and better-informed guidance for clients. It also holds the promise of preemptive advice to avoid risk, which is almost always more cost effective than litigating the effects after the fact. New legal services based on analytics may also increase the size of the legal market and help firms win new clients.
If firms can automate some of the most time-consuming tasks of providing legal services, they can provide the services at lower cost and can afford to help many more clients. At lower cost and higher scale, law firms can afford to serve many more clients, including those without the means comfortably to hire a lawyer but who nevertheless do not qualify for assistance from legal aid.
The American Bar Foundation’s now-famous 2014 Accessing Justice study showed that some 80 percent of people with legal problems don’t address them through the legal system. The majority of people who handle legal matters themselves fare less well than people who have the benefit of counsel. This means that the vast majority of people with a legal problem are disadvantaged because they do not or cannot avail themselves of legal counsel, suggesting a latent market for legal services.
The hope of access-to-justice advocates is that law firms can use AI, data analytics and automation to deliver newer and better legal services at lower prices to more people. In 2017 Microsoft announced it was partnering with Pro Bono Net and the Legal Services Corporation to develop an open-source access-to-justice portal to help people connect to justice resources or lawyers who can help them. Pro Bono Net has also used rules-based AI tools from Neota Logic to help lean-staffed legal aid clinics triage calls, visit websites and help requests.
AI is not just a way to reach the latent market. Some of the earliest firms building AI tools serve large corporate clients. Companies have been using data analytics and AI for all kinds of business tasks. In essence, everything except their legal spend.
Clients are demanding that law firms become more sophisticated about understanding their own costs. In 2017 Microsoft announced that it was shifting from paying its law firm lawyers hourly rates. So now law firms are using AI and data analytics to better understand their own costs, thereby allowing corporate clients to create more accurate legal budgets in advance.
Equally important to clients, when law firms bid fixed fees on their engagements, they become highly incentivized to work efficiently. Law firms still have a long way to go to understand staffing and efficiency, but new tools are helping them to understand their costs. For example, Digitory Legal uses AI to understand past bills to help law firms create accurate litigation budgets, a task that was previously considered impossible.
How Law Firms Use AI (Really)
The revolution of using AI in law firms isn’t coming — it’s already here. But it’s the next step in this revolution that is most promising. Law firms that use AI built into a third-party tool, or analytics reports from legal research services, have access to insights that those who don’t use the tools won’t have. Using third-party tools gives advantages over using no tools at all. Indeed, reporting analytics and AI tools should be the new floor. Law firms that do not use AI or analytics tools will scarcely be able to compete for new business.
In fact, lawyers have been using AI in their practices for years. Every time Microsoft Word autocorrects a spelling error, it’s using AI. Computer scientists often muse that “once it works, we stop calling it AI.”
That may be why lawyers look past the AI tools they use in practice. Electronic discovery (or technology-assisted review) now has higher accuracy and recall than human reviewers, using supervised learning. Lawyers also use third-party tools like RAVN from iManage to categorize and summarize documents or LawGeex to ensure the consistency of contracts.
LegalMation has built an upload tool using IBM Watson components that analyzes complaints (for specific causes of action and jurisdictions to start), creating a first draft of an answer with fact-specific defenses as well as discovery requests, within about two minutes. The documents aren’t quite ready to file in court, but they can save days or weeks of time preparing the first draft.
Casetext, ROSS Intelligence, v|lex, and Judicata have launched tools that use AI to analyze briefs, helping to identify missing cases or the strongest arguments. Fastcase built the world’s first algorithmic citator, Bad Law Bot, which identifies cases that have been overturned using text analysis and citation analysis tools.
The insights garnered by these third-party tools will be a kind of standard. Any firm that uses the same tools will likely achieve the same insights. Of course, lawyers can achieve better outcomes from different interpretation of the data, different training or unique strategic decisions based on data.
The ceiling for legal work will be defined by the custom tools that law firms are just starting to build. Take as an example the knowledge management systems that larger firms have built over the last 10 years. These systems were designed to collect and leverage the combined wisdom of firm lawyers over the years, thereby providing noncommodity insights for later matters. Similarly, the AI tools of the next few years will leverage the private data of law firms to create unique insights unattainable by other law firms because they are generated using the collective experience of lawyers and their work product from a particular firm.
New Tools: From Read Only to Read/Write AI
Law firms are becoming laboratories of experimentation in AI tools, spawning a generation of “makers” in the legal services industry. Katherine Lowry, director of Practice Services at BakerHostetler, recently upgraded the way lawyers at the firm asked questions of the knowledge desk. She embedded many repetitive questions into a Microsoft chatbots platform that resembles chatting with a human assistant.
It isn’t too much of a stretch to imagine legal aid clinics similarly helping to resolve similar legal aid inquiries that don’t involve legal advice.
Law firms are using AI to dig deep into their matter histories to understand the resources required to handle different kinds of client matters. Clients are increasingly requesting fixed-fee engagements or alternative fee agreements from law firms. But if those firms do not understand their costs, a fixed-fee engagement poses a serious risk of cost overruns borne by the firm.
So instead of hand coding and curating past bills, firms are using AI to understand the range and distribution of costs, computing the mean and median costs for similar matters and looking for facts that create outlier conditions. Understanding costs mitigates risk for clients and for law firms, and it can help those firms be more competitive when seeking new business.
Similarly, law firms are looking at litigation analytics more than ever to determine litigation strategy — in no small part because the tools are better. Tools from the recently launched Lexis Analytics and from Docket Alarm give a deeper look than ever at the strategies, judges and law firms that help firms to understand litigation outcomes. Formerly the domain of federal courts only, these tools are now expanding into state courts as well.
Markets may drive this trend toward a deeper understanding of legal analytics. Clients need better information to make strategic decisions about litigation, and they are becoming increasingly sophisticated about pricing risk. In addition, litigation financing companies will have hundreds of millions of dollars at stake, so they will demand that firms are using analytics to understand the risks at trial.
Expert witnesses have similarly been a mystery to many litigators for years; understanding the strengths or weaknesses of different experts, or their challenge histories, is a difficulty that faces firms large and small. Litigators are now using AI tools such as Fastcase’s AI Sandbox to combine their own experience with those of expert witnesses, then factoring in expert witness data from Courtroom Insight, to create their own custom dossiers for expert witnesses.
How Law Firms Become Makers
How is it possible for law firms, long reputed to be technological laggards, to build their own tools for pricing risk or driving litigation strategy? The trend is powered by the same forces that have more generally driven the advance of AI.
For the first time there is more legal data than ever to train machines. Law firms have vast document repositories and legal research databases are less expensive than ever. There is more information available from email and the internet generally for firms to use.
Second, processors continue to scale according to Moore’s Law, doubling in processing power and halving in size and cost every two years. Computing power is now a commodity, and law firms can more easily use cloud computing platforms such as Amazon Web Services, Microsoft Azure or Google Cloud to massively process data at scale.
Third, the AI tools are better and easier to use than ever. We think of IBM Watson as a monolith that beat Ken Jennings and Brad Rutter at “Jeopardy.” But Watson is more properly a collection of discrete pieces of software, or application programming interfaces, that accomplish specific tasks. Today you can use more than 60 of IBM’s Watson Developer Cloud APIs and get started with many of them for free. Instead of the six-figure price tag for AI of old, Watson application programming interfaces now cost fractions of a penny per call. (But be careful, those penny fractions can add up!)
Lastly, there is more open-source AI software than ever. TensorFlow from Google is perhaps the best known in this genre of open-source AI software. But last year LexPredict released the first open-sourced AI tool specifically for lawyers, ContraxSuite. The ContraxSuite platform can be used to analyze documents, extract and organize information and visualize the data for clearer understanding.
An AI Future for Law?
Will law firms adopt AI? Remember that 20 short years ago, people were seriously asking whether lawyers would ever use the internet.
We are just at the leading edge of the AI revolution in law, but we are not waiting for it to begin. Law firms are already using AI tools, and now they are using AI to build new tools for themselves. This is not a revolution for the next generation; it’s an important opportunity for ours.
Perhaps it is easiest to think about what will happen if firms don’t use data analytics and AI. For one, their peer firms are already starting to use AI, and that literacy makes those firms more competitive for client business. Firms that do not learn to use these tools will become less competitive.
This is not a revolution for the next generation; it’s an important opportunity for ours.
If their law firms do not use AI to deliver legal services, corporate legal departments will take more of the legal analysis work in-house. In 2016 corporate legal departments brought in-house $4 billion worth of work from outside counsel, a trend we can expect to continue if sophisticated clients outpace their law firms in analytic understanding of their legal spend. And finally, if law firms do not want to help corporate legal departments use analytics and AI, the Big Four accounting firms and consulting firms would be happy to rush in to fill the void.
Law firms have a great opportunity working with the next generation of data analytics tools and AI. They can reach latent markets for legal assistance, develop new and deeper insights for sophisticated corporate clients and gain a differentiated strategic advantage over trailing firms.
Ed Walters is the CEO of Fastcase and vice chair of the Board of Pro Bono Net. He teaches The Law of Robots at the Georgetown University Law Center and at Cornell Tech in New York City. He serves on the editorial board of RAIL: The Journal of Robotics, Artificial Intelligence & Law, and he is the editor of Data-Driven Law(Taylor & Francis 2018). A version of this article originally appeared in the January/February edition of Law Practice Magazine.