OTTAWA — Canada’s Artificial Intelligence Minister Evan Solomon said Monday he has summoned senior staff at OpenAI to discuss “safety protocols” after the tech company decided against reporting a Canadian ChatGPT user who police say went on to kill eight people in a school shooting.
Solomon said he will meet with OpenAI’s senior safety team on Tuesday in Ottawa after speaking with them over the phone on Sunday.
“We will have a sit-down meeting to have an explanation of their safety protocols and when they escalate and their thresholds of escalation to police, so we have a better understanding of what’s happening and what they do,” Solomon told reporters.
OpenAI had banned the account of Jesse Van Rootselaar seven months before police said the 18-year-old killed eight people, including five children, in Tumbler Ridge, B.C. on Feb. 10. The Wall Street Journal first reported that Van Rootselaar’s ChatGPT account was internally flagged after some employees interpreted her writings as “an indication of potential real-world violence,” and urged company leaders to alert Canadian police.
An OpenAI spokesperson told The Wall Street Journal that at the time of the banning, Van Rootselaar’s account activity didn’t meet the company’s criteria for reporting to authorities. The company reached out to police after they learned of the shooting, she said.
“Those reports were deeply disturbing, reports saying that OpenAI did not contact law enforcement in a timely manner,” Solomon said.
OpenAI did not immediately respond to POLITICO’s request for comment, and it was not immediately clear who from the company will meet with Solomon. But it appears the company agreed to meet with Canadian officials because neither members of Mark Carney’s government, nor Parliament, can summon a person who lives outside of Canada. However, it can enforce the summons if they ever set foot in the country, a move that would be considered extremely rare.
Canada’s Justice Minister Sean Fraser, Public Safety Minister Gary Anandasangaree and Culture Minister Marc Miller are also involved in the investigation.
Solomon wouldn’t say how far the Canadian government is willing to regulate AI chatbots or if this incident reshapes their online harms strategy.
“I’m not going to pre-judge. The details of this case, obviously I can’t get into, but I will say this … we are making sure that all options are on the table to make sure that Canadians are kept safe,” Solomon said.
Canada’s Liberal government has promised an online harms bill for five years, but has struggled to strike the right balance between protecting children online while preserving online speech. Two iterations of the bill failed to pass through Parliament under then-Prime Minister Justin Trudeau.