JAY GOLDBERG: Expanded Online Harms Act to regulate AI would be mistake

Article content

Rumours are swirling around Ottawa that the Carney government will release a new version of the Trudeau government’s Online Harms Act later this spring.

Article content

Particularly in the wake of the Tumbler Ridge tragedy, Carney government ministers have speculated a new version of the Online Harms Act could be expanded to cover liability for AI chatbots.

Article content

Recommended Videos

Article content

This shouldn’t be a surprise, especially given how aggressively cabinet ministers like artificial intelligence minister Evan Solomon have been pushing companies like OpenAI to be more proactive in reporting risks identified in private chats to law enforcement. But an expanded version of the Online Harms Act would represent a ticking time bomb when it comes to Canadians’ privacy rights.

Stiff opposition

The previous version of the Online Harms Act faced strong opposition in Parliament, particularly from the Conservatives, for not being limited enough in scope. Before the legislation died when Prime Minister Mark Carney called the 2025 election, the government had narrowed it to some degree to exempt private communications and proactive monitoring from the legislation.

If a new piece of legislation were expanded to include regulating AI chatbots, both private communications and proactive monitoring would be brought back into the bill. In other words, any progress made in making the old Online Harms Act less invasive when it comes to personal privacy would be thrown out the window.

Article content

Article content

Even with the changes the Liberal government made, the previous Online Harms Act still went too far in how it planned to regulate the use of social media platforms. Throwing private communications and proactive monitoring into the mix would take the bill from bad to worse.

Previously, Section 6(1) expressly provided the Act’s duties do not apply to any private messaging feature of a regulated service. That would have to be taken out, or fundamentally altered, to govern private communications, which would open up a whole can of worms. That’s because, as Professor Michael Geist of the University of Ottawa put it, “Chatbot interactions align far more closely with private messaging than social media because they involve one-to-one exchanges rather than public distribution.”

Do we really want to end up in a situation where the government is passing laws to govern private messaging, either on social media or with AI chatbots? That’s a slippery slope that Ottawa should very much avoid.

Sweeping new legislation, including reviving and expanding the Online Harms Act, is not a sound way to deal with the Tumbler Ridge tragedy. Government should work with platforms like OpenAI and Anthropic to ensure their reporting mechanisms are improved. But sweeping new legislation is not the answer.

Article content

We cannot forget the Tumbler Ridge tragedy was a failure on multiple fronts. The shooter, Jesse Van Rootselaar, was already known to the authorities. There were previous visits to the home by police, forced hospitalizations, and gun confiscation, among others. Before the government introduces an expanded Online Harms Act to deal with this issue, which would threaten the privacy rights of all Canadians, it must be remembered that Van Rootselaar was long on authorities’ radar.

New territory

When it comes to regulating AI, we are clearly entering new territory. That new territory demands smart rules, not sweeping new legal obligations that Canadians shouldn’t have to tolerate.

Under an expanded Online Harms Act, users would likely have to be worried about getting flagged and potentially reported to the authorities because new standards would be both overly broad and too low. Canadians don’t want virtual cops on their devices. And if regulating private conversations comes back into the equation, it’s not just AI chatbot interactions that would be regulated, but private social media messages as well.

Article content

The Online Harms Act ought to be left in the past. Canadians cannot allow the Carney government to use a tragedy that happened in Tumbler Ridge to revive deeply flawed legislation and make it worse. Politicians need to balance both public safety and civil liberties. Reviving and expanding the Online Harms Act, far from representing a balanced approach, would be a massive encroachment on Canadians’ civil liberties. And that cannot be permitted.

Jay Goldberg is the Canadian Affairs Manager at the Consumer Choice Center

Article content