AI Visionary Avi Gesser is building incredible solutions to challenges in cybersecurity, privacy, e-discovery, and other legal matters. We were thrilled with his perspective on how far AI has come, and how it can help teams in all areas of business do more, better.
The legal sector has a reputation for being slow to embrace new technologies, but you stand out as an early adopter of AI. What are some of the structural barriers that keep law firms from adopting new technologies? How and why did you take an interest in AI?
There are several cultural factors at work here. Some lawyers are just not very technologically curious or are creatures of habit. Lawyers are busy, and if we find a way to get something done that works, we can be reluctant to make significant changes. There is also the fact that most lawyers think of themselves as selling their time, so we may be uncomfortable with technologies that package up our services more like a product.
I got interested in AI as a prosecutor. We were conducting an investigation in which we had collected over a million documents. We were looking for communications on a particular issue, but everyone we were looking at was using different terms and acronyms to describe that issue, so search terms were only getting us some of the relevant documents. We experimented with a very early version of predictive coding, and it showed the promise of AI to help lawyers quickly find and organize the most important documents when they are buried in a giant sea of data.
Can you share any lessons on how legal leaders can drive the adoption of new technologies such as AI inside law firms?
What’s worked for me is finding the right tool, starting small, and making incremental improvements over time. Law firms are extremely good at generating useful work product for particular clients for particular matters. But we are not always great at building on that work product across different matters for different clients.
Finding the right tool to harness all the amazing work that is already being done, and channeling it to a constantly improving set of organized data, is very challenging—but totally worth the effort.
For instance, to help us track all the regulatory developments in AI, our team has created a simple database: the Debevoise AI Regulatory Tracker, or DART. The DART takes new pronouncements or guidance on AI and breaks it down into dozens of subcategories (e.g., governance, explainability, vendor management, bias testing, employee training, etc.). That allows our team to easily assess areas of convergence, trends, and outliers—all of which we use to help our clients predict where AI regulation is heading. The idea is to enable them to build models today that can withstand AI regulatory scrutiny in the future.
You have had an illustrious career—from working on some landmark cases during your time at the Department of Justice, to more recently working with blue-chip clients on cybersecurity matters. Looking back, what professional wins are you most proud of and why?
I’m not sure this counts as a “win,” but I am most proud of the work that our team did as part of the Department of Justice’s BP Oil Spill Task Force. Prosecutors often face criticism for not taking on hard cases. That was a very hard case. It was extremely complicated legally, factually, and technically. But it was a righteous case, and the team spent more than two years working tirelessly in New Orleans, away from their families, to get the job done.
What were your interests early on and what drew you to the practice of law?
My father was a professor of chemistry and physics, and my older brother is an engineer, so I grew up in a science-oriented house. I studied chemistry and math in college but found the lab work isolating, and the push toward increasingly narrow fields of expertise didn’t really appeal to me. On the other hand, the social aspect of the law, and the constant innovation it requires, were very exciting. In math, there are few truly novel developments—but in law, we create entirely new subject areas all the time.
What do you do when you are not working? How do you decompress?
Lately, I’ve been playing a lot of ping pong.
Which person (living or deceased) do you most admire?
I clerked for U.S. District Court Judge I. Leo Glasser, who is still on the bench at age 97. He is an extraordinary public servant and human being. He received a Bronze Star for bravery in World War II. Upon coming home, he enrolled in Brooklyn Law School, where he taught for many years, eventually becoming dean. He was then a family law judge until being appointed as a federal District Court judge in 1981. He has presided over several high-profile organized crime cases, including the trial of John Gotti. Judge Glasser is a model of dignity, compassion, integrity, and justice.
What do you consider your greatest achievement?
In 1982, I was very briefly the Rubik’s cube champion of Winnipeg, Canada.
Which historical figure do you most identify with?
I have always greatly admired the writer Martin Gardner, who spent most of his life explaining science concepts to non-scientists. He wrote wonderful puzzle books for both children and adults and creatively used basic physics and chemistry to become one of the most important magicians of the twentieth century, and perhaps the only one who regularly explained how his tricks were done.
What do you consider the most underrated quality or skill?
Curiosity. Pretty much all of my more-favorable qualities come from being a generally curious person.