This week, I am attending Summit AI New York. Summit AI is a global conference and exhibition focused on the application of AI in business in general as opposed to legal in specific.

As I have mentioned before, I like attending nonlegal tech conferences because they often yield insights we don’t get in our legal tech cocoon. That was certainly true from the opening Summit AI Keynote this morning.

The Keynote was a fireside chat with Matthew Fraser, the CTO for New York City. I almost didn’t attend the Keynote since I figured it would not yield anything possibly relevant to the law. But I was wrong.

New York’s Use of AI

Fraser talked a lot about how the City used AI and Gen AI to help people in small but doable ways. The City uses AI to help people and small businesses obtain benefits and navigate the regulatory system.

 “AI is not just about the technical benefits of AI; it’s what you can do from a human perspective,” said Fraser. As part of this approach, the City first recognized that it had to do such things as provide broadband access for everyone in public housing. Broadband access is essential for so many things, like enabling kids to do homework that needs to be done online.

This access enabled the City to then use AI to help residents find the benefits to which they may be entitled. Fraser noted that a person may, for example, have a vague notion they might be entitled to a certain benefit but not what other benefits to which a person in a similar situation typicaly is also entitled. He gave the example of someone who might be entitled to subsidized housing but who would not know they would also be eligible for food and health benefits. The City uses AI to notify people in a peer group that they are preapproved for certain benefits by just providing a certain amount of information, much like preapproved credit cards.

Fraser said the City also recognized that small businesses faced challenges navigating a complex regulatory web. The City uses AI to help small businesses meet these challenges.

So What?

You may be wondering what does all this have to do with legal. Fraser told us a story about how the City introduced a chatbot that was hallucinating and providing wrong information. There was a hue and cry that the bot should just be taken down. Instead, Fraser’s group went to work trying to fix the bot and eliminate the bad information. Within a short period of time, the bot was doing much better. Said Fraser, “When you stumble, it’s important not to stop, try not to look at it as a stop sign, but as a setback to be overcome.”

I thought about what New York is doing and Fraser’s story and how that relates to much of the discussion about access to justice. All too often, we focus on how AI is not doing much to reduce the A2J problem. We look at traditional legal problems and wring our hands about how AI probably wont move the needle at al on these big challengesl. We fail to focus on what problems AI can solve and the need to supply building blocks to enable more significant gains later, much like New York realized it had to first insource broadband access if it wanted to use AI to help people.

New York hasn’t let what Gen AI can’t yet do prevent the City from using Gen AI to solve the problems it can.

What New York is really doing is attacking A2J challenges in practical ways that can actually help people meet day to day problems . New York’s chatbots may not solve more complex legal problems of the kind we typically focus on like criminal justice challenges. Challenges where today a lawyer is needed to solve. But the New York programs are a start in using AI and Gen AI to solve the problems the platforms can solve. New York hasn’t let what Gen AI can’t yet do prevent the City from using Gen AI to solve the problems it can.New York looked at problems people have for which a lawyer is not needed but for which lay people nevertheless have trouble solving, like untangling benefit eligibility knots.

It’s human nature, I suppose, to focus on the problems and lose sight of the fact that AI can indeed impact A2J problems, but perhaps not in the grandiose and immediate way we think it should. Yes, the City has perhaps taken small steps in solving the overall A2J problem. But small steps are at least steps and may lead to larger steps later.

Perhaps we should applaud the small steps and think about how the lessons learned can lead to next steps that do more.