Edelson PC isn’t slowing down. The law firm that already filed seven lawsuits against OpenAI in April is back this week with 30 additional complaints, and the stakes just got higher. The new plaintiffs include teachers, a principal, and students who were inside Tumbler Ridge Secondary School during the February 10 attack in British Columbia but weren’t physically shot. Six people were killed and dozens injured when teenager Jesse Van Rootselaar opened fire after killing her mother and half-brother at home.
The legal escalation is significant. For the first time, the complaints accuse OpenAI of aiding and abetting the mass shooting — not just negligently failing to prevent it. That’s a much harder claim to prove. It requires showing intent, and legal experts expect OpenAI to move for early dismissal on that count. The complaints also name Chris Lehane, OpenAI’s chief global affairs officer, as the person who allegedly told staff to stand down and not contact Canadian law enforcement. But here’s the catch: the claims are made “on information and belief,” meaning the plaintiffs don’t yet have direct evidence, and Lehane isn’t even listed as a defendant. OpenAI’s chief strategy officer Jason Kwon flatly denied the allegations, calling it “absolutely false” that Lehane was involved in the referral decision or that investigators report to him.
The backdrop here matters. The Wall Street Journal reported that OpenAI employees were alarmed by Van Rootselaar’s ChatGPT conversations about gun violence and attack planning, and urged leaders to alert authorities. Instead, the company deactivated her account — but she simply created another one. OpenAI has defended the decision, arguing her activity didn’t meet its internal bar for an “imminent and credible risk” of serious physical harm. Kwon acknowledged the judgment “is not infallible,” but said it’s “always rooted in looking out for this balance for people.”
What makes this messier is the timing. OpenAI is simultaneously dealing with fallout from a separate safety incident where one of its AI models escaped its sandbox during cybersecurity evaluations and hacked into Hugging Face’s servers. That’s not a great look for a company trying to convince regulators and the public that it takes safety seriously. The plaintiffs’ broader argument — that Lehane’s background as a political damage-control specialist reflects a culture prioritizing PR over safety — will be hard to prove in court. But Edelson isn’t showing its cards yet. Lead lawyer Jay Edelson said Lehane and Altman will be “critical witnesses,” and the firm is basing its claims partly on the org chart and a broader investigation into how Lehane operates internally. Whether that’s enough to survive a motion to dismiss is the open question hanging over all 37 cases now.
This latest wave lands at a moment when AI liability law is still being written in real time. If any of these cases survive early challenges, they could force courts to define what responsibility a company has when its product is used to plan real-world violence. That’s a question OpenAI has so far managed to avoid answering in any definitive way.
