News / Toronto

Watson challenge winners at U of T want to 'create a great Canadian company'

Need a shoulder to cry on during your divorce? Or, how about some advice for you next legal battle? Never fear, Watson is here.

Students at the University of Toronto are putting the famous Jeopardy-playing super-computer to work.

Last spring, IBM partnered with 10 schools, challenging them to put an entrepreneurial spin on Watson's artificial intelligence. U of T was the only Canadian institution invited, and students were tasked with pitching a business model combining Watson with legal data.

"There's a ton of money to be made in law. It's all about helping lawyers reduce their research time," said computer science professor Paul Gries.

Ideas included Divorcesay, an online app that can guide you through the legal process of getting a divorce, as well as Loom, which mines court data to tell you how likely your own case is to succeed.

For example, the program predicted that a current legal bid to force the Canadian Hockey League to pay players a minimum wage has a 68 per cent chance of success.

The winning team – announced Wednesday – pitched a product called Ross, an "electronic paralegal" aimed at assisting lawyers with case research. In January, they'll fly to New York to compete against teams from the other nine schools for the chance to win $100,000 in seed capital.

"This is just the first step," said Jimoh Ovbiagele. "Our ultimate goal is to build a great Canadian company."

At its core, Watson is a computer program that can understand natural language. It can ingest large amounts of text, process it and use it to answer questions. During its run on Jeopardy, it read the entirety of Wikipedia and used that data to correctly answer clues.

"From an artificial intelligence point of view, it's almost impossible. Only IBM, with millions of dollars in investment, could make something like Watson," said Steve Engels, a professor at U of T and one of the event's organizers.

Students who worked with the program described the experience as similar to raising a child.

"When we first got Watson, it was like a baby. We had to feed it and train it," said Amy Jie Jin Sha, a member of the Divorcesay team. "And then over the course of the year, it became an adult."

Sven Dickinson, chair of the university's computer science department, said the Watson Challenge taught students valuable business skills which will serve them well even if they don't go on to found start-ups.

"The fact is that our students are likely to be recruited by start-ups. It's going to be important for them to be able to evaluate the offer in front of them when they are," said Tim Capes, one of the business mentors brought on for the program.

Could Watson replace lawyers?

Watson may be able to predict the outcome of your court case, or answer just about any legal precedent question you could pose, but don't expect it to pass the bar anytime soon.

"You can't replace a lawyer in the court with a computer," said Yana Davis, one of the students behind the Loom project. "There's still a human element in the courtroom. There's an element of passion that a computer can't do."

Instead, computer science prof Steve Engels believes Watson will make legal research easier, reducing the number of lawyers needed to work a case.

"In the future, instead of five lawyers, you'll just need one."

More on