Canada’s Minister of Artificial Intelligence, Evan Solomon, is demanding greater transparency from OpenAI following revelations that warning signs linked to the Tumbler Ridge, B.C., mass shooting were not escalated to police before the attack occurred.
The company behind ChatGPT acknowledged this week that it had failed to refer the shooter’s prior activity to law enforcement, despite having identified violent behavior on the platform months earlier. The admission has prompted federal officials to consider whether new legislation may be needed to regulate how artificial intelligence companies respond to credible threats of violence.
In a statement Friday, Solomon said OpenAI’s public commitment to improve its safety protocols lacked detail. “Canadians deserve clarity,” he said, emphasizing the need to understand how decisions about human review, escalation thresholds, and privacy considerations are made when potential threats emerge online.
Solomon confirmed he will meet with OpenAI CEO Sam Altman next week to seek assurances that concrete measures will be implemented. He also plans discussions with other AI firms in the coming weeks to ensure a consistent national approach to reporting and youth protection.
“All options remain on the table,” Solomon said, signaling that legislative changes could follow if voluntary reforms fall short. “Public safety must come first.”
The scrutiny comes after OpenAI disclosed that, under its updated internal policies, the shooter’s activity would likely have been flagged to authorities had it been detected today. However, at the time, the content was not referred to police until after the deadly attack in Tumbler Ridge, which left eight people dead and dozens injured.
Further complicating the issue, the company revealed it discovered a second ChatGPT account associated with the suspect after she had been identified by police — despite her first account having been suspended months earlier for violent content. A system designed to detect repeat violators had failed to prevent the creation of the new account.
OpenAI has since pledged to strengthen both its police referral processes and its mechanisms for identifying repeat offenders. The company also promised to establish direct points of contact with Canadian authorities and improve efforts to connect users displaying troubling behavior with local mental health supports.
Federal ministers expressed frustration earlier this week after OpenAI did not present a detailed action plan during a meeting in Ottawa. They indicated they would give the company an opportunity to outline specific steps before deciding whether formal regulation is required.
Calls for stricter oversight have grown louder in the wake of the tragedy. Researchers and opposition Members of Parliament have urged the government to accelerate plans to regulate artificial intelligence technologies, citing gaps in accountability and reporting standards.
British Columbia Premier David Eby also signaled he will seek stronger commitments from OpenAI. While acknowledging that the company ultimately provided information to law enforcement, Eby described the failure to escalate earlier warnings as a “colossal, horrific mistake.”
He said he intends to press Altman for support on a national reporting standard that would require AI companies operating in Canada to alert authorities when users appear to be planning violence. Such a framework, he suggested, should include clear consequences for non-compliance.
As federal and provincial leaders prepare to meet with OpenAI executives, the broader debate over AI governance in Canada has intensified. The Tumbler Ridge tragedy has become a stark reminder of the complex balance between privacy, innovation, and public safety in an era increasingly shaped by artificial intelligence.
