Law Technology Today From the ABA Legal Technology Resource Center
Related Articles
Artificial intelligence is set to change everything. Over the next few years, it will transform every aspect of the legal business. But many lawyers still underestimate its impact, and many law firms haven’t grasped the huge opportunity it presents.
Right now, legal AI lags behind artificial intelligence in other sectors, such as retail and finance. But this won’t last because the industry needs AI to cope with today’s new demands. The only thing holding legal AI back is a lack of source material for it to learn from.
AI Will Be Bigger Than You Think
There is a misconception that AI won’t change very much in law. Many lawyers still think that AI is just another piece of bolt-on software that won’t make much difference to our daily lives. Yes, it will speed a few things up; automate more processes and do a few extra steps our current software can’t manage. But essentially, it’s going to be business as usual.
This is wrong. AI is much bigger than that. The real power of AI lies in its creativity. In fact, if you think the power of AI lies in automating more things, you have the whole idea upside down: the power of AI comes not from our ability to tell it what do to, but from its power to tell us what to do.
How will this work in practice? In the future, AI will actively suggest new legal ideas for us to consider. For example, artificial intelligence—after collating, synthesizing, and analyzing information about corporate law—could suggest entirely new entities, such as more tax-efficient cross-border structures, for us to consider.
That Is Because AI Is Really Intelligent
How is this possible? How can computers suggest totally original ideas? Ideas that have never been seen before? That is because true AI doesn’t just simply mimic human behavior—it learns from that behavior and applies those lessons in new contexts. When it applies these lessons in new scenarios, the power of AI can truly astound us. I personally call these “Move 37” moments.
Over the last few years, Google’s DeepMind has been developing AlphaGo, a highly advanced piece of AI, to play Go, a difficult strategy board game. Last year, the AI went up against Lee Sedol, one of the world’s best players.
Not only did the AI win a series of games, but even more remarkable was a certain move AlphaGo made during the game: to be precise, Move 37 in game two. The move was a shock because although AlphaGo’s AI had been trained on millions of games stretching back through human history, this move was utterly unique—no one had ever played this move before, or anything like it. AlphaGo had somehow transcended its source materials to figure out something new, fresh, and unique.
How had it done this? Although AlphaGo had been trained on historic games, the developers at Google had then set up the AI to play against itself, over and over again. This allowed it not only to apply the lessons it had learned from analyzing these millions of games, but also apply them in new contexts, and learn yet more things.
A combination of learning from source material, and applying that knowledge in new ways allowed AlphaGo to transcend human-level playing ability—and defeat one of the world’s best players.
The Lessons We Can Learn In Law
So, AI has the potential to surprise us with Move 37 moments—but only if it has the source material to learn from. And this is the missing piece of the jigsaw that will enable AI to take the legal industry to the next level. Currently, legal AI doesn’t have enough information to learn from. It needs to be able to see, watch, and learn how lawyers use and apply their knowledge.
Half the battle is already won. AI can already access, organize, and synthesize primary legislation. It knows what the law is—but it doesn’t yet know how to apply that law. How can AI access this additional information? It will only be able to learn what it needs by “watching” lawyers at work; observing them discuss, argue, and debate legislation; see how they interpret terms and apply them; read their interactions to understand the way lawyers think and the logic they use.
We could create a new platform where lawyers, their clients, and regulators can discuss, annotate, debate, ask questions, provide answers, and generally talk about legislation. This will create a social layer of data on top of raw legal documentation. It is this social data that AI will be able to use to learn from: not only understanding what the law says, but how it is applied. This is, in fact, what we are creating right now at Waymark Tech.
Legal AI will transform the way that law works. In the future, legal AI will go beyond just automating processes, and, instead, actively suggest new ideas and provide guidance on interpreting and applying laws. The benefits for law firms are clear: new ideas, approaches, and methodologies for clients.
But to get there we have to create this layer of social data. And to do that, as an industry, we need to be open-minded about collaborating and sharing insights, both with our colleagues and AI.
