by Kathleen Finnerty, William Kellerman and Derek Ulmer
Kathleen Finnerty, a business litigator and counselor in practice for 25 years, William (“Bil”) Kellermann, a technology litigator and counselor for the past two decades, and Derek Ulmer, a 2017 graduate of Pacific McGeorge, share their views about Artificial Intelligence (AI) – how it is already affecting the practice of law and what to expect in the future. After discussing how law practice has changed (a lot) since they started practicing decades ago, Kathi and Bil turned to Derek’s questions.
Ulmer: Let’s start with some examples of how AI is already deployed.
Finnerty: Both ROSS (by ROSS Intelligence) and CARA (by Casetext) are “smart” legal research tools, meaning they get smarter with experiential data. For example, counsel can drag and drop a brief into CARA and within seconds have all relevant case law, both good and bad, in a folder for review, with insights on the most relevant cases, including those previously cited by the lawyers in other cases. Other AI can be used to review standard contracts, in milliseconds, to identify deviations from previously used forms.
AI is also being used for dispute resolution. For example, eBay has a “Resolution Center” available to resolve disputes between buyers and sellers which are typically too small to take to court. EBay reports that – win or lose – users are pleased to have a means to address their grievances.
Kellermann: Clients now have access to AI programs that help them identify a suitable lawyer in a community where they expect to litigate. To identify an attorney with a particular expertise, AI systems (such as Sky Analytics) mine data from on-line sources, such as Pacer and LinkedIn, to paint a compelling picture of why a client should hire a specific lawyer, regardless of firm size or location. Well positioned local firms that optimize their on-line presence, just as marketers optimize search engine hits, can now “fight above their weight.”
Ulmer: In broad terms, with automation soon performing many of the routine tasks involved in practicing law, what skill sets will attorneys – especially younger attorneys – need in the future to remain relevant and to provide value to their employers and clients?
Finnerty: Lawyers will need to know something about how AI operates. Within our varied law practice settings, we should know enough both to explain to our colleagues how AI operates and to check the validity of our AI’s work-product. We will also need to understand enough to defend our AI work-product when it gets challenged, and enough to help devise strategies to challenge our opponents’ AI work-product. That means we will need to know something about how certain AI systems are programmed, what assumptions are built into their operation, and whether any of those assumptions can be challenged.
Kellermann: Lawyers, and especially litigators, will need to have a basic understanding of statistics to know if a system’s output makes sense in light of the input. With systems using psychometrics to influence choices, all lawyers must understand newly developed theories in cognitive science – when, for example, an argument or line of questioning is based on triggering confirmation bias. This skill will be necessary to exercise discretion and judgment in direct advocacy, in challenging the argument of the opposition, and in challenging the inputs or outputs of an AI system.
Ulmer: Recent news stories speculate AI will soon perform the lesser tasks historically performed by junior attorneys. As a recent graduate, should I be worried?
Kellermann: Worried? No. Those who understand, deploy, and master these new tools will always have a valuable, marketable skillset. Be curious – learn how to use new tools and, at a basic level, how those tools work (including both their strengths and limitations). When dealing with AI outputs, be a skeptic in a scientific way – trust but verify. These new technologies can be daunting because they promise to change how we practice. But they are also exciting for the efficiencies and enhanced results they will provide for clients.
Finnerty: The days of young lawyers pouring over volumes of depositions and dredging through thousands of documents are waning. Young lawyers are increasingly tasked with overseeing the use of technology, such as assisting in designing search criteria at the front end and then validating the AI program’s work-product at the back end, as opposed to conducting granular document-by-document reviews. What will always be necessary is critical thinking to challenge AI’s results, and creative thinking to develop and achieve new solutions. The deployment of this technology, combined with corporate clients’ aversion to paying for young lawyers’ time (and training), will surely impact how young lawyers will obtain legal training and mentoring.
Kellermann: Andrew McAfee at MIT studies the effect of AI on society. He discusses the term “HiPPO” – Highest Paid Person’s Opinion. His studies reveal that a HiPPO’s decisions are correct only 8 percent of the time, that AI alone is correct better than half the time, but most importantly, a HiPPO decision supported by predictive data is correct almost all the time. AI systems are a powerful new tool, but their deployment will still require the same human skills lawyers have always provided: judgment, discretion, advocacy, skepticism, creativity. While a computer program may identify all the relevant documents better than people, it will still take human judgment to validate the machine’s work-product, or to find the 200-300 “hot” documents necessary to win a summary judgment motion or at trial.