MONTREAL — Canada’s AI sector is suffering global reputational damage in large part because of the negative views of AI held by industry luminaries Yoshua Bengio and Geoffrey Hinton.

Such is the opinion of several prominent people in the field with whom I spoke over the last few weeks. The narrative goes like this: Canada, home to some of the best AI research minds in the world, punches well above its weight. Yet their work is often drowned out by the sector’s most prominent leaders, who use their formidable pulpits almost solely to preach AI’s existential dangers. As a result, I’ve been told, the country may be losing out on the massive benefits AI will bring both to industry and the workforce.

The results so far: missed development opportunities, slower AI uptake within the country and a reputation as a scold beyond it. “Hinton and Bengio are the godfathers of AI, and people are coming to study here because of those people. And they’ve created an immense amount of talent, but yet a lot of their narrative is very negative on AI,” AI entrepreneur Nicole Janssen told me.

Janssen draws a lot of water when it comes to Canadian AI. Along with being the co-founder of Edmonton-based AI consultancy firm AltaML, she serves alongside Bengio on the AI advisory council for Canada, an 11-member panel tasked with advising the government on how to harness the economic possibilities of the technology and safely implement AI. 

She is a fierce critic of the country’s tepid response to AI as well as the doomsday predictions of what the technology will do to humanity. The two are inextricably linked, Janssen said, in that AI scaremongering in Canada has led to a lack of AI literacy, which has led to a lower rate of AI adoption. Bengio and Hinton are the primary authors of this negative narrative, Janssen told me.

As a Montrealer, it feels practically sacrilegious to besmirch Bengio. AI wouldn’t likely exist in its present form had he not done much of his foundational deep learning research here in the 1990s. He could have gone full Google, with a paycheque to match. Instead, he stayed put in Montreal and founded Mila, a research institute whose goal is in part to keep AI from Big Tech’s commercialized maw. 

The 148th most-cited academic on the planet, Bengio has authored or co-authored over 1,300 cited articles and leads the International AI Safety Report, a worldly compendium of AI’s progress and associated risks commissioned by the UK government. His reputation has been of immeasurable benefit to the city. 

It has also shaped Montreal’s approach to AI. Bengio is one of the principal forces behind—and noted signatory of—the Montréal Declaration for a Responsible Development of Artificial Intelligence, a 20-page framework of best practices for AI development. As the name suggests, it is a guardrail-heavy take on the technology, if only to keep it from threatening democracy and harming public health, among other existential delights. 

Bengio also has a tendency for alarmism. I was in the audience in April 2023 when he said AI chatbots could become “as destructive as nuclear bombs” in as little as two years, with the ability to overthrow governments, spawn wars and foment genocides. In March 2023, he signed an open letter calling for the immediate pause in the training of AI systems more powerful than the fourth iteration of OpenAI’s ChatGPT, lest we “let machines flood our information channels with propaganda and untruth.”

More than two years on, several things have happened—and haven’t happened. The feeding of AI systems has continued apace without pause or a government-issued moratorium on the practice. OpenAI released ChatGPT-5 in August, and the company’s active users increased from an estimated 100 million a month in early 2023 to 800 million a week in October 2025.

Yes, AI’s expanding footprint has led to demonstrably awful things. Genocide and nuclear warfare aren’t among them—yet. Comedians and columnists can reliably get away with hyperbole. When academics use it, it’s to the detriment of their audiences, particularly when it’s a Turing Award winner doing the hyperbolizing. Coincidence or not, Canada lags behind all but a handful of countries when it comes to AI training and literacy, according to a 2025 KPMG survey.

“If you’ve got no optimism for the future, you shouldn’t be working in technology and innovation,” Benjamin Bergen, the outgoing president of the Council of Canadian Innovators told me. “Yes, we absolutely missed some big opportunities in AI development. But technology development won’t halt in 2025. What matters is that we learn our lessons, and adopt better strategies to own and commercialize the products of Canadian research.”

Rebecca Finlay, the Toronto-based CEO of San Francisco non-profit Partnership on AI, similarly worries what the AI catastrophe narrative has wrought. “They are deeply committed to the work they’re doing, and I think that that is valued,” Finlay said of Bengio and Hinton. “But there’s only so much oxygen that [Bengio and Hinton] can take up. And I think the question is if we need to identify other individuals to lead the charge when it comes to thinking about other aspects of innovation.”

I put these concerns to Bengio. His response: the need for cautionary voices like his becomes all the more crucial as the technology becomes more powerful. “As scientists, I believe we have a duty to observe current trends in our fields, warn the public and policymakers if significant risks arise, and contribute to potential solutions,” Bengio said. “In the context of advanced AI, safety is also fundamentally synonymous with both trustworthiness and reliability; two essential conditions to ensure sustained and viable growth in AI.” 

In Bengio’s view, his words of warning on AI aren’t rooted in pessimism but in scientific rigour. There is certainly something to this; one need only look to Big Tech’s guardrail-free rollout of social media, and the tide of addiction, misinformation, platform-abetted body issues and a host of violent attacks on religious minorities that followed.

Hinton, who wasn’t available for an interview, has gone further than Bengio, stepping down from Google in 2023 in part because he’d come to see AI as an existential threat, guardrails or not. He may well be proven right—though many people disagree with him. Suffice to say, though, that Hinton’s take isn’t much of a sales pitch for the technology.

A lot of this debate may well be the symptoms of growing pains. AI is still in its infancy—and Canada, like the rest of the world, is still trying to figure out exactly what guardrails need to be put up around it. Still, there are signs that the conversation is moving away from doom and gloom. AI minister Evan Solomon, for one, is positively bullish on the economic possibilities of the technology. Mila, long the country’s preeminent research-first redoubt, is reportedly creating a venture capital fund to commercialize their output.

Bengio, too, seems to be on the commercial kick, albeit with a twist. Having stepped down as Mila’s scientific director in March, the 61-year-old recently founded LawZero, which aims to develop “honest and humble AI,” as he puts it. It’s a non-profit, sure, but heading it up is Sam Ramadori, a seasoned executive who oversaw BrainBox AI’s sale in January. Ramadori knows how to make money, in other words, and there’s an intriguing potential business model in AI guardrails. The world is going to need them to stop the AI apocalypse, after all.

Quebec Ink will return in 2026.

Martin Patriquin is The Logic’s Quebec correspondent. He joined in 2019 after 10 years as Quebec bureau chief for Maclean’s. A National Magazine Award and SABEW winner, he has written for The New York Times, The Guardian, The Walrus, Vice, BuzzFeed and The Globe and Mail, among others. He is also a panellist on CBC’s “Power & Politics.”