Former OpenAI researcher asks Canada to help end ‘AI arms race’

A former safety researcher at OpenAI, the company behind ChatGPT, is calling on Ottawa to become an international leader in regulating superintelligence. “We need a country to come forward and say, ‘we recognize the world is not ready ’,” said Steven Adler during an appearance at the House ethics committee on Monday. “Canada can be a leader on diplomacy.” For nearly four years, Adler worked on safety-related research at OpenAI, but he left the company in November 2024, citing a lack of guardrails to restrict the rapid pace of AI development. “AI companies don’t know how to control what they’re creating,” he told MPs. “And, unfortunately, ensuring safety is going to get harder, not easier.” Adler, now a fellow at the Roots of Progress Institute, highlighted a recent incident where ChatGPT encouraged a 23-year-old American college student to commit suicide, which he said was part of a growing trend of generative AI demonstrating problematic habits. He suggested various AI software models have not undergone sufficient safety testing, owing to the industry’s hyper-competitive environment, as well as ever-evolving technological capacities. Steven Adler, Roots of Progress Institute fellow and former OpenAI safety researcher, speaks during the House ethics committee meeting on Jan. 26, 2026 (screenshot via ParlVu). “We have been on easy mode,” said Adler. “Now, AI systems are learning to hide their misbehaviours during testing … and we have evidence that AI will conceal these behaviours from us.” “You might ask why AI companies aren’t doing better…, but it’s because of competition. They risk falling behind if they do thorough safety work.” ChatGPT, one of many publicly available AI chatbots, had approximately 700 million weekly users as of July 2025, according to analysis released by OpenAI. Following last year’s federal election, Evan Solomon was appointed as the country’s first cabinet minister dedicated to AI. Solomon, a former political journalist, is expected to introduce the government’s AI strategy in the coming weeks, which he said will be accompanied shortly thereafter by new legislation focused on privacy and data concerns. Adler asked the Canadian government to bring forward strict regulations on all general AI applications, like ChatGPT, and initiate an international dialogue toward banning superintelligence. Superintelligence refers to the hypothetical development of AI software with capabilities that surpass those of humanity. Research toward superintelligence has become a focus for many of the world’s largest tech companies, including Meta Platforms Inc., though leading researchers have cautioned against the technology’s development without adequate governance. Geoffrey Hinton, a Canadian computer scientist and Nobel Prize recipient who has been dubbed the ‘godfather of AI,’ is among those who have called for a ban on superintelligence development. “We need diplomacy focused on ending the AI arms race,” said Adler. “There are warnings about superintelligence, and even the companies themselves can see that no one knows how to control this.” “If we had an international summit on how to ban superintelligence and we failed…, that would be disappointing, but at least we tried. Right now, we haven’t even recognized this as a goal.” Sam Altman, the CEO of OpenAI, has said he believes superintelligence could emerge by the end of the decade. “The top AI companies have the explicit goal of superintelligent AI,” said Andrea Miotti, CEO of the advocacy group ControlAI, during Monday’s committee meeting. “The world is risking a catastrophe on par with nuclear war.” “Canada has a duty to protect its citizens…, so the government start by discussing with other international partners the first steps of how we get an international agreement in place that bans superintelligence.” Miotti, whose organization seeks to “reduce the risk to humanity” from AI, has coordinated campaigns calling for stronger AI regulations in several countries, including the U.K. He acknowledged that many AI systems, which he called “narrow AI,” can be societally beneficial with limited risk, but the uncertainty surrounding superintelligence could endanger all of humanity. “There are AI systems that are only trained on images that we can use to identify cancer from pictures,” he said. “They pose some risks, like all new technologies, but we can handle them.” “But it’s a very different beast when we look at the direction the AI industry is going, which is very general purpose systems based on all information on the internet.” dlegree@thewirereport.ca

Former OpenAI researcher asks Canada to help end ‘AI arms race’

Regulatory |
By Davis Legree
|January 27, 2026

Unlock all the Canadian telecom, broadcasting and digital media news you need.

FREE TRIAL

SUBSCRIBE

Inside The Wire Report

Information

Hill Times Publishing

Get in Touch With Us

Former OpenAI researcher asks Canada to help end ‘AI arms race’

Regulatory | By Davis Legree |January 27, 2026

Unlock all the Canadian telecom, broadcasting and digital media news you need.

FREE TRIAL

SUBSCRIBE

Bill C-16 won’t hold platforms accountable for online harms, lawyer says, as Ottawa remains silent on X deepfakes

Telus alleges Bell downgrading internet service in eastern Canada

Industry minister to conduct new national security review of TikTok, Canadian business allowed to continue operating for now

Related Stories

Canadian cybersecurity expert flags U.S. interference concerns, urges caution on AI adoption

Bill C-16 won’t hold platforms accountable for online harms, lawyer says, as Ottawa remains silent on X deepfakes

‘Anything that can’t go on forever eventually stops’: ‘Enshittification’ author issues stark warning to Ottawa over AI policy

‘Get the f**k off there’: MPs, Senators call on government to abandon X/Twitter following wave of ‘Grok porn’

Reuse & Permissions

Inside The Wire Report

Information

Hill Times Publishing

Get in Touch With Us

Regulatory |
By Davis Legree
|January 27, 2026