OpenAI Hit With 30 Suits Over ChatGPT Safety Flag and Ban Loophole

For AI and machine learning practitioners, the Tumbler Ridge litigation is less a legal story than an engineering and trust & safety postmortem: an automated review system flagged conversations about gun violence, human safety staff recommended notifying authorities, and a high-level executive allegedly declined for reputational reasons. Worse, the enforcement action was reportedly a deactivation rather than a system-wide ban, allowing the user to re-register with another email.

On September 2, 2026, OpenAI and chief executive Sam Altman were hit with 30 new lawsuits in a California federal court, accusing the AI developer of aiding and abetting the suspect in Canada's Tumbler Ridge school shooting. The complaints were filed by students, teachers, and the principal who were present at the school, and they allege that OpenAI gave "substantial assistance and encouragement" to alleged shooter Jesse Van Rootselaar. This is the second major wave of litigation tied to the attack, following suits brought in April 2026 by victims' families. The new filings add operational specificity to a theory of liability that, if accepted, could redefine what AI platform providers must do once they know a user is escalating toward violence.

“

On September 2, 2026, OpenAI and chief executive Sam Altman were hit with 30 new lawsuits in a California federal court, accusing the AI developer of aiding and abetting the suspect in Canada's Tumbler Ridge school shooting.

At the center of the new complaints is an alleged sequence of internal safety failures. The plaintiffs claim OpenAI's automated review system flagged the suspect's ChatGPT conversations about gun violence, and that the company's safety team recommended contacting Canadian authorities. Yet according to the suits, OpenAI's chief global affairs officer, Chris Lehane, was involved in a decision to remain silent because executives worried about the company's "reputational and financial standing." Instead of a system-wide ban, the complaints say the platform "deactivated" Van Rootselaar's account, which allowed him to regain access by signing up with another email. The lawsuits assert that providing the shooter with continued access to the tool constituted substantial assistance in itself and was a key factor in planning and carrying out the attack. These are serious, fact-intensive allegations that put OpenAI's internal decision-making and safety governance directly at issue.

OpenAI's public response has been forceful. Chief strategy officer Jason Kwon posted on X that the lawsuits contain "false claims" about the safety teams. He added that it is "completely untrue to say that the people at the center of these challenging decisions do not prioritize safety, or that there are 'political' or 'public relations' factors at play." The statement defends the company's culture but does not directly address the specific allegation that a deactivation, rather than a ban, created a re-registration loophole. Meanwhile, the Tumbler Ridge filings are not the only legal front: OpenAI is also facing a lawsuit from the state of Florida, which accuses the company of aiding and abetting mass shooters, including the person who carried out an attack at Florida State University last year.

From a legal and regulatory perspective, these cases sit at the intersection of product liability, platform immunity, and AI safety. Early internet liability battles focused on whether platforms had to remove harmful third-party content. The Tumbler Ridge theory is different: it targets the design and operation of a generative AI product, how OpenAI built its automated review, what it did after a safety flag, and whether its enforcement actions were meaningful. If courts allow the aiding-and-abetting or substantial-assistance claims to proceed, AI developers may face a duty not merely to moderate outputs but to act on risk signals, escalate to law enforcement, and enforce bans in a way that prevents easy re-registration.