ILTACON 2016 Session: Machine Learning as a Service and the Future of AI and the Law
Daniel Katz, associate professor of Law at Chicago Kent College of Law, opened his Day One ILTACON session by assuring attendees that the days of robot overlords are still a bit further off.
“I’m not about robots replacing lawyers, I’m about placing robots where they aren’t now,” Katz remarked.
Lawyers, technologists and futurists have long been captivated by the promise of AI in the law. As Katz noted, looking at AI in the law needs a reset. In many ways, rules-based AI – or the use of decision trees to guide professional and non-professional users with information – exists today with at-home tax preparation tools.
As Katz described, AI beyond these uses is often “very brittle,” but in a world where driverless cars are a reality, the promise of more complex systems are right around the corner. The difference is that computer systems can now collect data to learn to “mimic or predict what a person will do.” As he noted, now we are merely looking at a big data problem.
To date, optimizing analytics has been more profitable in medicine, finance, logistics and agriculture. According to Katz, the legal industry may be a more “narrow vertical,” but it is incredibly useful, particularly in e-discovery.
As Katz noted, big data and analytics is often a frightening space for lawyers, but it has been deployed in many ways, like predictive coding, transactional work and more. He noted that predictions can come from three sources: experts, crowds and algorithms.
“We need to evaluate experts and somehow benchmark their expertise,” Katz said, adding that frequent Fantasy SCOTUS winner Jacob Berlove, who is not an attorney, has figured out how to harness this ability. As Katz described, the future of AI in the law is about finding “a way for the signal to beat the noise.”
As he assured attendees, being an expert, or just having an algorithm, alone isn’t enough, the real power is having both. This certainly doesn’t look like a future worth fearing.
“The democratization of machine learning is on the way,” Katz said.