Why the world isn't ready for robot lawyers

This month the consumer credit rating agency Equifax finally admitted it had been hacked over the summer, releasing the personal data of millions of their customers. People have been understandably angry.

To fight back, a student at Stanford University named Josh Browder created a bot to prepare the required documents to sue Equifax without having to go through a lawyer. This 'robot lawyer' was an extension of his previous project, DoNotPay, which allowed people to fight traffic tickets automatically.

But not everyone thinks the world is ready for robot lawyers. Ryan Calo is a law professor who focuses on issues of artificial intelligence and robotics. Just because a bot can challenge a traffic ticket doesn't mean it's equipped to take on a huge company. "If you challenge a parking ticket, often the state or the local municipality will back off. It's not very often that you have to litigate that up in court. But if you're talking about filing a lawsuit for thousands of dollars against a company like Equifax, they're going to push back and then your bot is going to run out of options."

"The notion that this is ultimately going to replace people — lawyers, other trained actors in the system — that strikes me as not just improbable, but even a little dangerous." Dangerous because, when it comes to the law, the stakes can be very high.

 "When it comes to law it's how we reconcile real conflict, and often people's property, even liberty, even their life can be on the line."

While bots might not be ready to take the place of lawyers completely, they are finding a role in the legal system.

Susan Wortzman is a partner at the law firm McCarthy Tétrault. She leads the firm's e-Discovery and information management practice. That means that, as part of her job, she frequently has thousands, even millions, or emails or digital records to look through for a case. So to cut down some of that work, she uses an artificially intelligent program that sifts through all those records to figure out what is important and what can be ignored.

"When I started practicing years ago it was two boxes of documents and that would take me a couple days to get through those two boxes. Now if I'm starting with two terabytes of documents, several million records, even using all the technology there's still going to be those key documents... that need to be reviewed. The complexity of the cases requires the lawyers working on it have to review those records. The reality is there's a lot more evidence now."

Their system is about 98% accurate, compared to 60-70% when this job is done by a human. But despite how successful this system is, Susan doesn't see it putting lawyers out of work.

"AI is doing all this work and it's making our job easier, but we still need the input, we still need the lawyers."