In the wake of the Tumbler Ridge mass shooting last month, many are asking the question of where the line should be drawn between public safety and personal privacy.  

The Tumbler Ridge shooter, Jesse Van Rootselaar, was banned from OpenAI back in June for troubling posts. Still, the contents of those posts were not initially reported to the authorities, despite internal debate.  

JAY GOLDBERG: Tumbler Ridge tragedy demands policy change, not sweeping new government powers Back to video

Following that news, Artificial Intelligence Minister Evan Solomon summoned senior officials from OpenAI to Ottawa but emerged from the meeting “disappointed” and seeking “concrete proposals” for the company to change its protocols.  

OpenAI said it had taken steps to update its policies, thanked the government for a “frank discussion” and said that dialogue with Ottawa would continue.  

Solomondeclared after the meeting that “all options for us are on the table, because at the end of the day, Canadians want to feel safe.”  

Weighing public safety with privacy concerns

The question of where to draw the line between public safety and personal privacy is one on the minds of global public policymakers as new technologies emerge.  

Where the line has traditionally been drawn is that there must be imminent danger and the possibility of active harm. The AI chat logs, made by Van Rootselaar back in the early summer, weren’t deemed to meet that threshold by OpenAI.  

By putting all options on the table, Prime Minister Mark Carney’s government seems prepared to consider changing that standard.  

But the government must be prudent. New duty of care rules, which would redefine when AI platforms must flag potential concerns, would require new legislation.  

While the Carney government might be looking to pursue just that, it must be careful not to increase legal liability or duty of care in a way that degrades the user experience or risks the privacy of Canadians.  

In the Tumbler Ridge case, it is clear that self-regulation by OpenAI wasn’t perfect. But Van Rootselaar was known to law enforcement as well. There were visits to the home by police, forced hospitalizations, and gun confiscation, among others.  

Before blaming technology, the government must examine the failings of other institutions.  

AI companies shouldn’t be saddled by overreaching legal obligations

And, when it comes to regulating AI, we are entering new territory that requires smart rules. We cannot impose new legal obligations that go beyond what would be tolerated in other industries.  

Users shouldn’t have to be worried about getting flagged and potentially reported to the authorities if new standards are too broad and have thresholds that are too low. Too broad would mean over-monitoring, over-filtering and over-reporting, creating virtual police on your device.  

We must remember that technology is a net positive for Canadians. AI companies are offering Canadians technologies that are already transforming the way we live, work, and play for the better.  

New rules meant to enhance public safety must, at the same time, be clear about not degrading the user experience nor curbing our privacy.  

Leave Online Harms Act in the past

And let’s be clear about another thing: The answer does not lie in reviving the Online Harms Act.  

The Liberal government continues to talk about reviving that legislation, which, as it was worded in the previous Parliament, already carries clear risks to free expression and digital privacy in Canada.    

This tragedy cannot be used to revive that flawed legislation and be sold to Canadians as a public safety measure.  

Politicians must balance both public safety and civil liberties. Smart rules and sweeping new powers for the government are not the same. The Tumbler Ridge tragedy demands action, but it must be thoughtful, measured and balance multiple principles that we hold dear.  

New targeted legislation may be warranted. Sweeping new powers, such as those proposed in the Online Harms Act, are not, and would do more harm than good.  

Jay Goldberg is the Canadian affairs manager at the Consumer Choice Center  

Eby says it looks like OpenAI could have prevented ’horrific’ Tumbler Ridge killings

RCMP confirm police contacted about Tumbler Ridge shooter's ChatGPT

Tumbler Ridge: Will we ever know shooter's motive?

© Edmonton Sun