Survivors and staff from the Tumbler Ridge, Canada school shooting filed 30 new lawsuits against OpenAI this week, one plaintiff per suit. For the first time, the complaints raise the allegation from negligence to aiding and abetting a mass shooting. Combined with the seven suits already filed in April, OpenAI now faces 37 separate claims tied to the same incident.
The shooting happened on February 10. Eighteen-year-old Jesse Van Rootselaar first killed her mother and half-brother at home, then opened fire at Tumbler Ridge Secondary School, killing eight people including herself and wounding dozens more. OpenAI has previously confirmed it identified and banned her account back in June 2025 for violating its usage policies, but determined internally that the activity didn't meet the threshold for referral to law enforcement.
The New Plaintiffs Weren't Shot
April's plaintiffs were families of the dead and people who were wounded. This time the 30 plaintiffs are a different mix: teachers, a principal, and students who were in the school building that day — including a 13-year-old girl, identified in the filing only as A.C., who survived by hiding in the library. None of them were shot; their claim is negligent infliction of emotional distress.
Adding this category of plaintiff expands the pool of potential claimants from the casualty list to everyone who was present that day — a completely different order of magnitude. The same firm, Edelson PC, filed both rounds, and the sequencing looks deliberate: establish the factual framework with the seven most severely affected plaintiffs first, then widen the exposure with a second, larger wave.
From Negligence to Aiding and Abetting
The new complaints list five causes of action: negligence, negligent entrustment, aiding and abetting a mass shooting, strict product liability, and negligent infliction of emotional distress. Most of these also appeared in April's filings; aiding and abetting is new.
The two claims sit at very different evidentiary bars. Negligence only requires showing the defendant should have foreseen the risk and failed to exercise reasonable care. Aiding and abetting requires evidence of actual knowledge and substantial assistance — closer to proving intent. Legal observers broadly expect this claim to face an early motion to dismiss. Plaintiffs' attorney Jay Edelson's response: "We're not laying out all our evidence at this stage," adding that Chris Lehane and Sam Altman would be "key witnesses."
The complaint names OpenAI Chief Global Affairs Officer Chris Lehane, alleging he told staff to stop briefing Canadian officials — while acknowledging it has no direct evidence to support the claim. OpenAI Chief Strategy Officer Jason Kwon's denial was blunt:
"It is absolutely false to say Chris Lehane was involved."
He added that describing people at the center of these decisions as failing to prioritize safety doesn't reflect the facts.
What They Actually Want Is the Chat Log
The plaintiffs' legal team's first discovery request is for the shooter's complete conversation history with ChatGPT, from her first use of the product through the day of the attack. The request carries weight beyond this case. OpenAI has consistently resisted turning over user conversations in litigation on privacy grounds, making the same argument in its copyright disputes with news publishers. The context here is different: the requesting parties are victims and families of the dead in a criminal case, and judges won't weigh that the same way.
After the shooting, OpenAI rolled out a round of product changes around minors and crisis intervention, adding a separate study-and-safety mode for teen accounts and revising its internal threat-escalation process. Under U.S. civil procedure, remedial measures taken after the fact generally can't be used directly to prove prior negligence, but they can shape how a jury perceives the company — a point not lost on either side's lawyers.
OpenAI's public position hasn't changed: it says it has zero tolerance for using its tools to facilitate violence and has strengthened its threat-assessment process. Edelson is speaking directly to the jury pool: "We're going to ask a jury to send OpenAI a clear message — that it can't make decisions that put profit ahead of children's lives."
Filing 37 separate individual suits instead of a class action is a deliberate strategy by the firm. Each case requires its own scheduling, discovery, and settlement negotiations, which stacks up the defendant's litigation costs case by case and makes it harder for any single global settlement to close out the exposure. For a company preparing for an IPO, that kind of uncertainty tends to sit in the risk-factors section of the prospectus for a long time.
Sources: TechCrunch, Global News, CocoLoop, CBC, CTV News; the lawsuit count, plaintiff composition and the five causes of action were cross-checked across multiple outlets, and OpenAI's statements are quoted from its executives' public responses.