Manitoba law professor studies Artificial Intelligence in courtroom

It’s prevalent technology that could shape the future of our Judicial system here in Canada, and it’s all because of extensive research by a University of Manitoba law professor exploring the impact of AI in the courtroom.

Dr. Katie Szilagyi studies law and technology, particularly investigating things about artificial intelligence and the use it has in the judicial system. She says there are multiple ways this technology is being utilized, including the use of automated decision makers.

For example — sentencing software — a tool already being used by judges through different statistical methods to assess how long a prisoner sentence should be.

Other ways include how likely someone is to re-offend and the use of facial recognition technology. But there’s one area in particular gaining traction.

“Natural language processing software. Things like ChatGPT, barred and different types of technology that are being used,” explained Dr. Szilagyi.

Szilagyi says AI and law have become enmeshed in the last little while and with the growth of AI industries, and it’s something we could see more of moving forward.

She says technology can be a very useful tool, but she gets worried when she sees others make claims that technology will be better than human expertise.

“Judicial sentencing tends to be fraught with bias and have charges of discrimination, misogyny, sexism, racism that are brought to the forefront by people in the social sciences. At the same time, we have to be critical at what we might be delegating those responses to.”

In June, Manitoba Chief Justice Glenn Joyal made the decision that attorneys in the province have to disclose if they have or plan on using AI to prepare court documents in the Court of King’s Bench.

“How (AI) will impact the justice system is anybody’s best guess,” said Chris Gamby, a Criminal Defence Lawyer with the Criminal Defence Lawyers Association of Canada.

Gamby says AI could hallucinate decision making within the court system.

“We operate on the common law principle in this country, meaning that there’s precedent. Like cases should be decided in a like fashion. If the AI is hypothetically inventing a precedent, that could obviously lead to some absurd results in the context of a preceding,” said Gamby.

Meanwhile, one of things that Szilagyi’s research really focuses on, is the storytelling aspect of law. She says it’s important to preserve human judgement.

“We have legacies of colonialism, discrimination, all sorts of things that show up in the data set because of the lived realities of the world we live over centuries and centuries. It’s very difficult to make sure that our data set is not dependent on those same things,” said Szilagyi.