Jason Boehmig - Law as Science, Science as Law - SuperAI Singapore 2025

All right, great to be here. I'm Jason, I'm a lawyer and I'm an entrepreneur. And today I'm going to talk about a shift that is happening in the legal profession towards a more data driven and a more scientific practice of law, and in particular how AI and dialogue with disciplines outside the law are helping shape that shift. So to do this, I'm going to look back at the writing of one of America's great jurists, writing over 100 years ago. And this is Oliver Wendell Holmes Jr. His life story, as you may be able to tell from looking at his desk, is very interesting. Before I went to law school, I worked for a couple years in investment banking, which I thought was tough and grueling. Before Oliver Wendell Holmes went to law school, he served in the Union army in the Civil War, where he was wounded three times, different time, and his legal career was distinguished. The New York Times has called him the second most notable US Supreme Court justice in history.

By name recognition. He's probably the first. His theories are still taught today. If you've heard the term clear and present danger, that's Holmes.

Now what I find most interesting about Holmes is the last writing that he ever did before joining the Supreme Court. In fact, this is his last extrajudicial writing he would ever do.

The topic was the beginning of a new way of thinking about the practice of law. He called it Law in Science and Science in Law. And the key argument that Holmes makes in this piece in his own words, is that the legal profession should make a substitution of quantitative for qualitative judgments. I'm going to repeat that because it's so important, a substitution of quantitative for or qualitative. Quantitative for qualitative judgments. Now, this was wildly futuristic in 1899.

To put things in context for you. When Holmes wrote this down, the modern field of statistics was 20 to 30 years away from what we would recognize, recognize it as today. The term computer meant a person with a ledger.

The core of his argument, though, is more relevant than ever. It is centered in his observation that the legal field operates in general experiential terms.

So to make this real, y' all have probably seen something similar to this and likely interacted with a lawyer that is working in general experiential terms in the commercial context at companies you work at. Here's what it looks like and I'm sure you've seen this before. It's usually accompanied by an email that says, well, we've never seen that before or we can't agree to that and A markup that looks like this. And all of that information comes from a lawyer's experience, from the algorithm that lives in their head. And sadly, this experiential practice of law is the norm. But Holmes tells us there is a different way. And because of AI and dialogue with other disciplines outside the law, that shift is happening today. For instance, at Ironclad, our legal team can tell you not just why they are agreeing to this clause, but what statistical effect agreeing to it has on our operations. Our general counsel can say, for instance, that conversion rate of our contract is increased by 5% when we shift to a more standard indemnification clause that conforms more closely with the version we frequently see from our vendors. So AI, obviously a big part of the shift in that data driven practice of law. And an insight like that. But so is another radical practice for lawyers talking to non lawyers. And in this case the lawn lawyers on our marketing team, they deal with metrics all the time, just like most business people. And can you imagine interviewing a marketer or hiring a marketer who says I don't believe in measuring the conversion rate of our website and instead only relied on their intuition for improving it? It would be pretty absurd, but that's the way most lawyers practice law today. And I think if we do our jobs right, and if Holmes vision comes to reality, it will be just as absurd for a lawyer to make a statement like that or work that way sometime in the near future.

Generative AI, obviously a particularly relevant school for the legal profession. It's good at working with semi structured text and as context windows get bigger, it gets better. We've seen what happens with computing capability when underlying technology continuously improves. And this would have excited Holmes because he saw the way towards a more scientific practice of law depends on having the right tools. Lawyers can't do this just on their own. And the way he articulates it is that any solution in general terms seems to me a mark to mark a want of analytic power. And what is AI if not a thousand fold increase in analytic power?

So I want to pose a question.

What would a similar prediction to Moore's Law look like for the output of lawyers? And I've kind of facetiously put down a new law here, that AI will dramatically reduce the time that it takes attorneys to produce the same work product. If we get anywhere close to this, it will have profound effects on the legal profession. For instance, Kirkland and Ellis, world's largest law firm, made 8.8 billion in total revenue last year, which is amazing. And to do that, they had to employ 3,828 attorneys. If something like this were to apply, Kirkland and Ellis would be able to achieve the same amount of revenue with 3.8 attorneys in 10 years from when this law starts to take effect. I actually had to, you know, I'm a lawyer because I had to pull up a calculator to make sure I was doing that calculation correctly.

And honestly, my actual views of how this is going to play out are a little bit more nuanced than that. But I think it illustrates the scale of change we're talking about when we talk about applying generative AI to the legal profession.

And I actually don't think this change stops with just the role of the lawyer. The more scientific practice of law will change the world of legal disputes of the type encountered by everyday citizens. This is the Magna Carta. And it's like the origin of the right to a jury of peers goes through the US Constitution and obviously is still around today.

But we all know kind of how unpredictable and imprecise peers can be. Holmes certainly did if you go read his full paper and one might rationally believe a future state of the world where an AI trained on one's peers is superior to those actual peers. If that sounds like science fiction, these are actual problems being worked on today by new startups, that there are AI arbitrators in production today, resolving disputes for consumers. There are AI contract negotiators, there are AI litigation associates. There are a really passionate group of people working to build the future of the profession. And I've been working in this field for about 14 years. This is by far the most exciting time of my career.

So one of the things that I appreciate about reading Holmes is he doesn't want to just throw away art as the profession moves to a science. His father was a poet and his vision is really that quantitative thinking can enhance our lives and enhance our appreciation of the work. So I love this quote comparing the qualitative casual listener to the technical quantitative listener. And the beauty to me isn't that the appreciation isn't lost on the technical listener listener. It's only enhanced. Knowing the chords and the keys doesn't mean you have to lose the tune. And that should be the way we think about AI in a legal profession. How we build in quantitative, scientific, data driven practices into an ancient profession without losing an appreciation of the art. So I'll leave you with that note. Thanks for listening. It's been fun to get to a chance to share this passion of mine. With you. And I guarantee you this is going to be the only article from 1899 from the Harvard Law Review you hear at Super AI and come find me if you want to talk about more about the history of law and how AI is shaping the future of the profession. Thank you.