The family of a girl critically injured in a mass shooting at a school in Canada has filed a civil lawsuit against OpenAI, alleging the company failed to notify authorities despite internal warnings about the suspect’s conversations with ChatGPT.
Twelve-year-old Maya Gebala was shot in the neck and head during the attack on 10 February in Tumbler Ridge, British Columbia, and remains in hospital with severe injuries.
Lawsuit allegations
The lawsuit was filed by Maya Gebala’s mother, Cia Edmonds. It claims OpenAI had prior knowledge that the suspect had discussed violent scenarios with its chatbot but did not alert Canadian law enforcement.
According to the filing, an account linked to the suspect, 18-year-old Jesse Van Rootselaar, had been banned by OpenAI in June 2025 because of the nature of the conversations conducted through ChatGPT.
The plaintiffs allege that several OpenAI employees flagged the exchanges as indicating an “imminent risk of serious harm to others” and recommended that police be notified.
Instead, the lawsuit claims, the request to contact authorities was rejected and the only action taken was the suspension of the account.
Claims regarding platform safeguards
The lawsuit further alleges that the suspect created the ChatGPT account before turning 18, which would require parental consent under the platform’s policies.
The plaintiffs claim that no age verification procedures were conducted.
It also states that the suspect treated the chatbot as a “trusted confidante” and discussed multiple scenarios involving gun violence over several days in late spring or early summer 2025.
After the first account was banned, the suspect allegedly created a second ChatGPT account and continued discussing violent scenarios.
According to the lawsuit, OpenAI “had specific knowledge of the shooter’s long-range planning of a mass casualty event” but “took no steps to act upon this knowledge”.
The shooting
The attack on 10 February is among the deadliest school shootings in Canadian history.
Eight people were killed, including five children and the suspect’s mother.
The lawsuit states that Maya Gebala was shot three times after attempting to lock a library door to prevent the shooter from entering. As a result, she suffered what the plaintiffs describe as a “catastrophic brain injury”.
OpenAI response
OpenAI told the BBC that the events represented an “unspeakable tragedy” and expressed sympathy for the victims and the community.
The company said it did not alert police at the time because the account did not meet its threshold for a credible or imminent threat of serious physical harm.
In recent weeks, the company said it has introduced new measures aimed at improving how potentially dangerous interactions are assessed.
These include the involvement of mental health and behavioural experts in evaluating cases and making the criteria for reporting risks to authorities more flexible.
OpenAI stated that under the revised guidelines, the suspect’s account would have been reported to law enforcement.
Government discussions and proposed changes
OpenAI Chief Executive Officer Sam Altman met virtually on 4 March with Canada’s Minister of Artificial Intelligence, Evan Solomon, and the Premier of British Columbia, David Eby.
According to the Wall Street Journal, Altman pledged to strengthen protocols for notifying police about potentially harmful interactions and to apologise to the community of Tumbler Ridge.
In an open letter to Canadian officials dated 26 February, OpenAI said it would also improve its detection systems to identify attempts to bypass platform safeguards.
The company added that it plans to establish a direct contact point with Canadian law enforcement to allow faster reporting of cases that may pose a risk of real-world violence.
Minister Solomon said that while the government welcomed the company’s willingness to improve its protocols, authorities had not yet received a detailed implementation plan.
Source: BBC