OTTAWA — Three years after the Competition Bureau began building a branch dedicated to applying competition law to the digital economy, the group is up to 45 people and has notched a win.

The tribunal that rules on competition cases found the Cineplex cinema chain broke the law by tacking on mandatory fees for digital tickets at the checkout stage, a form of “drip pricing” that’s been illegal since 2022.

“We had a behavioural economist who talked about the decision-making process on devices and how that works, and how human tendencies can be used to drive people to make certain decisions,” says competition commissioner Matthew Boswell. Cineplex is appealing, but the tribunal penalized the company $38.9 million.

The bureau is a watchdog assigned to keep companies that sell things from ganging up on their customers—one historically staffed by lawyers like Boswell, who is a former Crown prosecutor, and economists. Besides enforcing a new part of the law, the Cineplex case demanded a new kind of expertise that the competition authorities expect to apply more and more often.

“We’ve realized that we need different skill sets in the modern digital economy,” Boswell says.

Besides behavioural economists and psychologists, the bureau’s digital economy branch has been hiring data scientists and engineers, MBAs and experts in how business leaders make decisions. They’ve also brought in “intelligence specialists” who fall into two categories: tacticians who (for instance) determine exactly what to look for as the bureau plans a raid, and strategic thinkers who monitor markets for new competition concerns.

AI-assisted deepfaked product endorsements are one such concern, falling under the Competition Act’s prohibition on deceptive marketing. Lying about the goods you sell is illegal, to be sure, but how does a regulator fight faked reviews at scale, when whole batches of them can be created in seconds and posted on e-commerce sites?

Then there are the artificial intelligence tools that can abet collusion on prices.

“The use of AI for things like price fixing—that, to me, is a huge issue that is coming at us,” Boswell says.

His U.S. counterparts have taken aim at RealPage, a firm that offers tools for landlords to set rents. The company rejects the premise vehemently, but the U.S. Department of Justice is arguing that by using private data from multiple clients to recommend how to maximize their profits—data that competitors couldn’t legally share with each other directly—RealPage is facilitating collusion.

Boswell’s imagination goes a step beyond, to a future in which AI advisors can communicate with each other.

“What really scares me is AI models being used by companies and the companies not per se directing the AI model, but saying to the AI model: ‘Maximize the price I’m going to get for my product,’” he says. “The AI model itself reaches out to competitor companies’ AI models, and they say, ‘The best way to maximize our profits is if everybody charges the highest price.’”

Here’s the catch. The Competition Act forbids “persons” to collude or conspire with competitors to fix prices. But an AI tool is not a person.

“If it was three executives getting into a smoke-filled room and saying, ‘Let’s all charge the highest price and nobody deviate,’ that would be a crime punishable by up to 14 years in jail and no maximum on the fine,” says Boswell. 

If those same executives stay in their own offices and remain ignorant of their AIs’ split-second strategy sessions, though, it’s not against the current law. Even if the C-suite types do know—or might know, or might suspect but don’t ask too many questions—lawbreaking could be very hard to prove.

“That’s the kind of future stuff that makes me lose a little sleep,” Boswell says.

Artificial intelligence in the here and now is an ever-larger part of the bureau’s business, a dramatic shift since Boswell was named to his post in 2019 (his appointment, following a two-year extension, now runs to February 2026).

The Competition Bureau dedicated its annual summit in September to AI issues in its sphere and produced a voluminous report on the takeaways; an underlying theme was that the bureau needs to know more about the technology, the companies that develop it and those that use it, while trying to protect those markets.

The hard part is doing it without stepping on the progress and innovation that has come, to a significant degree, from some of the biggest companies in the world—firms that can afford the often-stunning costs of developing AI models, and seek to control the limited supplies of the tools needed to do it.

“One of the big concerns, globally, is the control of those key inputs already by Big Tech,” Boswell says. “Data, compute power, chips, cloud expertise, and money.”

The bureau has led a new forum of Canadian regulators to compare notes, especially on AI, that includes the Office of the Privacy Commissioner, the Canadian Radio-television and Telecommunications Commission and the Copyright Board.

One thing they’re all interested in is “synthetic content” like AI-created mimicry of real musicians’ work. How it must be labelled so as not to deceive audiences is one Competition Bureau question; what control people have over fakes that look or sound like them is one for the privacy commissioner; whether a fake Drake track counts as Canadian content is one for the CRTC; who controls the rights to AI-made work is one for the Copyright Board.

Cooperation among regulators is essential when dealing with new developments as disruptive as these, Boswell said.

“It’s common sense that people whose mandates intersect the way ours do get together,” he said.

Despite the enormous challenges, AI can help these enforcers, too, by sifting masses of data, looking for patterns, picking out what’s important.

“On many of our cases, we get hundreds of thousands, if not millions, of records that we have to analyze,” Boswell says. “There was a time when a person was going through each record, and you can imagine the kind of delay that builds into an investigation.”

The Competition Bureau’s hard cases end up in front of the Competition Tribunal—which is led by a judge—and often in judicial reviews after that. That means its work gets scrutinized and tested by lawyers working for companies that might have billions of dollars on the line.

“Our work is always going to be evidence-based,” Boswell says. “We always have to prove it in court.”