Brandeis professor develops tool to help laid off Canadians determine severance

Professor Jonathan Touboul and his collaborators in Canada used artificial intelligence to predict monetary awards in employment disputes. They've now made their program available to the public.

Judge's gavel hitting a sound block

Over the past several years, associate professor of mathematics Jonathan Touboul and several of his colleagues in Canada have used artificial intelligence to analyze court rulings in Canada's labor law cases.

They developed a series of algorithms that could predict the amount of termination pay a laid-off employee would get if she took her case to court before a judge.

With nearly two million Canadians out of work due to the COVID-19 pandemic, Touboul and his collaborators saw an opportunity to help.

In late May, Touboul, in cooperation with the Conflict Analytics Lab at Queen’s University in Ontario, launched the website MyOpenCourt.

After answering a series of questions, users are told roughly how much money they're owed in severance pay and connected with free legal consultation.

"We realized this was a moment when people needed access to legal information about their employment rights and the ability to pursue justice more than ever," Touboul said. "We decided, 'Let's get our research out there and available to the public. Let's go for it.'"

Canadian laws are generally much more protective of workers' rights than laws in the United States. Most employees are entitled to severance when they are let go.

For most workers, severance is determined by what are called the "Bardal factors."

These include the employee's age, the duration of their employment, whether or not they were managers, their experience and qualifications, and how easy it will be for them to find a similar job elsewhere.

When disagreements occur between an employer and employee over how to define one or more of the Bardal factors, one of the parties can take the other to court. Since 1960, when the Bardal factors ruling was first promulgated, there have been 1,500 labor law cases in Canada.

A team of volunteers working under Touboul and his two main collaborators in Canada — Queen's University assistant professor Samuel Dahan and McGill University associate professor Maxime Cohen — read through all of these cases and extracted information on the Bardal factors and the judge's ruling.

They then applied AI to look for patterns and correlations in the data.

In a paper posted online on June 4 and to appear in the McGill Law Journal, Touboul and his colleagues reported that they could roughly predict the amount of money an employee would win if she took her employer to court.

The prediction was accurate to within two months' worth of the worker's salary. Touboul said the variance was due to unique or uncommon factors in some cases that made the judges' decisions outliers.

Since it was posted, MyOpenCourt has had roughly 5,000 visitors. In addition to helping employees determine how much severance they're owed, it also helps them determine if they are legally considered a contractor or employee.

This has become an increasingly thorny legal question in the gig economy where, for example, someone may work as a driver for Uber, but not be considered an Uber employee and entitled to benefits.

When the site determines the user has a case, they are connected for a consultation with a pro bono labor law specialist.

Touboul, Dahan and Cohen plan to turn their attention next to personal injury law and consumer disputes both in Canada and the United States. Their objective is helping people determine whether it would be worthwhile taking their claim to court, especially those individuals who would find it very costly to hire a lawyer.

As it says on the MyOpenCourt website, "Everyone should have access to basic legal support, just like a living wage and health care."

In addition to Touboul and Dahan, the other authors of the McGill Law Journal paper are Jason Lam of Queen's University and Dan Sfedj of Brandeis.

Categories: General, Humanities and Social Sciences, Research

Return to the BrandeisNOW homepage