Edelson PC, the law firm that filed seven lawsuits against OpenAI in April on behalf of victims and families involved in the Tumbler Ridge shooting, is filing 30 more charges this week. The new plaintiffs include a teacher, principal and students who were in the building during the attack but were not struck by gunfire.
The new filing also charges OpenAI for the first time with aiding and abetting a mass shooting, rather than simply negligently failing to stop it. This claim would require proof of OpenAI’s intent and would likely face early rejection.
The complaint comes as OpenAI continues to respond to the aftermath of another high-profile safety incident, after one of its AI models escaped its sandbox during a cybersecurity assessment and was hacked into the servers of Hugging Face, a company that hosts open source AI models and datasets. OpenAI also faces several lawsuits alleging that ChatGPT’s design contributed to user suicide, acts of violence, and severe mental health crises.
A new wave of lawsuits filed Tuesday in California courts is related to the Feb. 10 school shooting in Tumbler Ridge, British Columbia, Canada. That day, teenager Jesse Van Rootseller murdered his mother and half-brother in their home, attended Tumbler Ridge Intermediate School, killed six more people, injured dozens more, and then killed himself.
The Wall Street Journal reported that OpenAI staff was alarmed by VanRootselaar’s use of ChatGPT, including conversations about gun violence and advice on how to plan attacks. OpenAI employees reportedly urged leaders to alert Canadian law enforcement about Van Loetzeler’s potential for violence in the real world, but leaders decided not to contact authorities. Instead, the company deactivated Ms. Van Loetzeler’s account, but she was able to quickly create another one.
OpenAI justified its actions by arguing that Van Rootselaar’s activities did not meet its internal standards for an “immediate and credible risk” of serious bodily harm to others, which is required for cooperation with law enforcement.
Jason Kwon, OpenAI’s chief strategy officer who oversees a team of human judges and the company’s legal department, said in a statement shared with TechCrunch that “its decisions are not absolute, but they are always rooted in a focus on balance for people.”
The new complaint specifically names Chris Lehane, OpenAI’s chief international affairs officer, as the person who instructed staff to resign and not contact authorities, but the complaint provides no direct evidence that he was personally involved, and OpenAI denies his involvement.
Mr. Lehane is a prominent political operator and PR spin master with damage control expertise on behalf of the Clinton administration, at Airbnb, and now at OpenAI. Plaintiffs argue that his background contributes to a culture at OpenAI that prioritizes public relations and damage prevention over safety.
One complaint shared with TechCrunch states that “the Intelligence Investigations Team, the only team within OpenAI responsible for identifying ChatGPT users who pose a threat of real-world violence, was placed under (Lehane’s) control.” “As a result, the decision to alert law enforcement to a user planning a large-scale attack was not made by the trained threat assessment experts who prompted OpenAI to contact (Canada Police). It was determined by Lehane himself, or someone in his chain of command, based on information and belief, and approved by Sam Altman.”
“On information and belief” is a legal term that means the plaintiff believes the claim to be true based on second-hand information, even if the plaintiff cannot directly prove it. TechCrunch was unable to confirm whether Lehane actually had the authority to override recommendations from the intelligence and investigation teams, or whether he was involved in this particular incident at all. Notably, Lehane is not named as a defendant in the complaint, but OpenAI CEO Sam Altman is listed as a defendant, as in the previous seven complaints.
“We haven’t revealed all the evidence at this point,” Jay Edelson, the lead attorney in the case, told TechCrunch, noting that Lehane and Altman will be “key witnesses” along with the safety team. “I can tell you that this is based in part on our org chart, but also on broader research, including what we learned about how Lehane operates internally.”
“It is simply untrue that Mr. Chris Lehane was involved in our initial referral decision or that our investigators were reporting to him in any way,” Kwon said. “It is also not entirely true to say that those at the center of these difficult decisions are not prioritizing safety, or that ‘political’ or ‘public relations’ factors are involved.”
The new complaint also seeks to rebut OpenAI’s previous defense that it did not alert Canadian authorities about Van Rootselaar’s conversations with ChatGPT, citing “imminence” and “privacy” justifications. The complaint points to an incident in November 2025 when OpenAI locked down its San Francisco office following threats from activists.
“While OpenAI acknowledged that there was no ‘indication of active threat activity’ and therefore no indication of an ‘imminent’ attack, the company…immediately locked down its offices, alerted employees, distributed the suspect’s name and photo, and notified the San Francisco Police Department,” the complaint states.
“When its own employees were in danger, OpenAI did not let ‘privacy’ grounds stop it from calling the police or distributing the man’s name and photo to thousands of its employees. It also did not wait when an attack was imminent because it knew waiting would put its employees’ safety at risk.”
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
