Nobel laureate and his former protégé headline public debate on AI at Toronto Tech Week.
Artificial intelligence (AI) godfather Geoffrey Hinton said big tech firms don’t want “regulations with teeth” before offering a pointed look to former protégé Nick Frosst—also the co-founder of leading Canadian AI company, Cohere.
“We invented this technology. Canada has every right to be a leader in it.”
Nick Frosst
Cohere
“AI big tech companies are like big oil companies” in their resistance to regulations, Hinton chided his former Google Brain employee.
It was one highlight in a series of public disagreements between teacher and student as part of a special Toronto Tech Week edition of the Desjardins Speaker Series put on by the University of Toronto.
Hinton, who is also the chief scientific adviser at the Toronto-based Vector Institute, has been awarded both the Nobel Prize in Physics and the Turing Award for his work on foundational principles of machine learning and deep neural networks. Frosst was one of Hinton’s first hires at the Google Brain lab in Toronto from 2016 to 2020.
Hinton’s lighthearted but pointed shot at big tech firms was far from the only source of contention between the Nobel Laureate and Frosst, whose company is Canada’s best-funded large-language model (LLM) developer, making AI products for enterprise.
Hinton said he believes that LLMs have subjective experiences and that neural networks, which he helped pioneer, work similarly to human brains.
RELATED: “Light, tight, right” regulation: Minister Evan Solomon unpacks how Canada plans to support domestic AI and quantum computing
Though he said it was “difficult” to disagree with a Nobel laureate on stage, Frosst took a different tack.
“LLMs are more conscious than a rock, but less conscious than a tree,” Frosst said, while Hinton argued that LLMs are in fact approaching human-level consciousness.
The discussion, moderated by CBC tech journalist Nora Young, touched on the short-term and potential existential perils of AI, which Hinton has repeatedly said will replace countless jobs and could lead to human extinction if left unchecked.
While Hinton warned that without proper guardrails, LLMs will invent new ways to scheme—something that fellow Turing Award winner Yoshua Bengio has said AI agents will do.
“They will be able to see creative new ways of finding people’s passwords,” Hinton said, noting that LLMs score well on creativity tests. One 2023 study published in the journal Nature found that, on average, LLMs outscore humans on divergent thinking tasks, but could not beat the most creative humans.
Frosst said LLMs hacking passwords was “unlikely,” adding that he and Hinton find themselves on “different sides of the spectrum” on the issue.
Hinton and Frosst also diverged in their outlooks on the state of AI development in Canada and the country’s place on the global stage.
“It will be very hard for Canada to stay at the forefront of AI,” Hinton said, adding that dedicated government investments such as the $2-billion Canadian Sovereign AI Compute Fund are helpful levers.
But the professor called Canadian businesses’ conservatism in adopting AI and educating their employees about it “a big problem.”
The conversation came a day after Cohere CEO Aidan Gomez encouraged domestic companies to resist any temptation to leave Canada on the Toronto Tech Week Homecoming stage, saying Canadian companies should instead build at home. Federal AI minister Evan Solomon also said AI regulations in Canada should be “light, tight” and “right” in remarks yesterday at Toronto Tech Week.
Frosst, whose company has signed a slew of large enterprise customers in recent months, including a memorandum of understanding with the federal government, was more bullish on Canada’s AI scene.
“We invented this technology,” Frosst said. “Canada has every right to be a leader in it.”
BetaKit is the official media partner of Toronto Tech Week. Feature image courtesy University of Toronto.