OpenAI is speaking out about changes and overhauls coming to its safety protocols and communications with law enforcement after a mass shooting in Tumbler Ridge, British Columbia, Canada, that left nine dead, including the shooter, earlier this month.
The perpetrator of the shooting had their ChatGPT account suspended in June 2025 after OpenAI detected content from the user that presented “an indication of potential real-world violence.” OpenAI banned the user and considered alerting authorities. However, OpenAI decided not to inform law enforcement after determining that the user had no credible plan to act out.
OpenAI’s decision not to alert authorities has become a major concern of the Canadian government.
While OpenAI did have a protocol to handle credible threats, the company is now saying it will do more. In an open letter to the Canadian government, OpenAI’s Vice President of Global Policy Ann M. O’Leary did not offer any specific policy changes, but did mention that changes were already being implemented and more were coming.
In addition, O’Leary also shared another oversight in OpenAI’s protocols. According to OpenAI’s open letter, the Tumbler Ridge shooter had opened a second ChatGPT account, which the company only discovered after the shooting occurred and the name of the shooter was publicly released. OpenAI did share that account with police after making the discovery.
Mashable Light Speed
In its open letter, OpenAI says it will “continue to strengthen” its law enforcement referral protocol. The company said it’s working with mental health experts to better assess difficult cases and when conversations cross the line into being a risk.
OpenAI also said they are establishing a direct point of contact with Canadian law enforcement to refer cases with potential for “real world violence.”
ChatGPT will also be tweaked to provide help to users who are in distress or are pursuing prohibited behavior by providing them resources to localized support within their communities.
In addition, OpenAI said it is addressing its protocols that allowed for the shooter to open a second account. The company said it had a system in place to detect repeat policy offenders and is committing to “strengthening our detection systems to better prevent attempts to evade our safeguards and prioritize identifying the highest risk offenders.”
Mashable has reached out to OpenAI for additional information regarding these policy overhauls and to find out whether these affect the company’s policies in the United States as well. We will update this piece when we hear back.
