New Lawsuits Filed Against OpenAI in U.S. Court Over Canadian School Shooting
The litigation tests whether a U.S. AI company can be held liable in American federal court for alleged failures to flag violent communications that preceded a mass casualty event in a foreign...
A teacher and additional students injured in a February 2026 Canadian school shooting filed new lawsuits against OpenAI in the U.S. District Court for the Northern District of California on September 2, 2026, according to a report by the New York Post. The new filings expand on seven earlier cases brought in April 2026 by families of victims killed in the shooting.
The April cases named OpenAI and its chief executive Sam Altman as defendants. Plaintiffs in those cases alleged that OpenAI prioritized profits over user safety, according to the New York Post report. The September filings add plaintiffs who survived the shooting but were injured in it, including at least one teacher.
The central allegation across all cases is that ChatGPT, OpenAI's conversational AI product, received messages from the suspected shooter that indicated plans for violence, and that OpenAI had an obligation to alert law enforcement before the attack occurred. The suspect has been identified in reporting as Jesse Van Rootselaar, 18 years old. The Post's account of the suspect's identity is based on its own reporting; court filings, once publicly available on PACER, the federal judiciary's case management system, would contain the operative complaint with the specific factual allegations.
The shooting itself took place in Canada in February 2026. Because the defendant, OpenAI, is headquartered in San Francisco, California, plaintiffs have chosen the Northern District of California as the venue. Whether a U.S. Federal court has jurisdiction over claims arising from a Canadian event is a threshold legal question that has not yet been resolved in these cases, as no published ruling on jurisdiction has been identified as of the date of this story.
The legal theory being advanced, that an AI platform bears a duty to report threatening communications to police, does not have settled precedent in U.S. Law. Courts have addressed the duties of internet platforms under Section 230 of the Communications Decency Act (47 U.S.C. Section 230), which provides immunity to platforms for content generated by third-party users. Whether that immunity applies to an AI system's interactive responses, rather than passive user-generated content, is a question that courts are only beginning to address. No final ruling has been issued in the existing April 2026 OpenAI cases that would resolve that question.
OpenAI had not issued a public statement in response to the September 2 filings as of the publication of this story. The company's prior public statements on safety, including its published usage policies and system card documentation, prohibit the use of its products to plan or facilitate violence. Whether internal systems flagged the alleged communications before the shooting is unknown. That question would most likely be addressed through discovery records in the Northern District of California cases, which are not yet public.
The April 2026 cases remain in early procedural stages. No docket numbers for the September 2 filings have been confirmed in publicly available court records as of this writing. PACER, the federal courts' public access system, would contain the operative complaints once they are entered into the docket.
The litigation arrives as Congress has held multiple hearings in the current session on AI safety regulation but has not passed a comprehensive federal AI liability statute. No binding federal law currently specifies what reporting obligations, if any, AI companies have when their systems receive communications that suggest imminent violence.
The Federal Trade Commission has previously investigated OpenAI over data privacy practices, opening a probe in 2023. That investigation did not produce a public enforcement action related to violent content reporting obligations. The FTC's jurisdiction over AI safety disclosures remains an open regulatory question.
What remains unknown in this story includes the specific text of the September 2 complaints, the assigned judge or judges in the new cases, whether OpenAI will move to dismiss on Section 230 or jurisdictional grounds, and what, if any, internal records OpenAI has regarding the alleged communications. Court filings on PACER under the Northern District of California docket, once entered, would answer the first two questions. A motion to dismiss briefing schedule, if one is set, would address the third.