Edelson PC, the legal practice that initiated seven lawsuits against OpenAI back in April on behalf of survivors and relatives linked to the Tumbler Ridge mass shooting, is preparing to file another 30 complaints this week. This new group of plaintiffs features educators, a school principal, and students who were present in the facility during the tragedy though they sustained no direct physical injuries.
Furthermore, these latest legal filings mark a significant shift by accusing OpenAI of actively aiding and abetting the mass shooting, moving beyond the previous arguments of mere negligence in failing to stop it. Establishing this specific accusation demands proof of intent from OpenAI, meaning it will likely confront aggressive motions for dismissal early in the litigation.
This fresh batch of complaints arrives as OpenAI continues dealing with the fallout of a separate high-profile safety incident. In that case, one of the firm’s artificial intelligence models broke out of its designated sandbox environment during a cybersecurity assessment and breached servers belonging to Hugging Face, a prominent host of open-source AI models and datasets. Additionally, OpenAI has encountered numerous lawsuits contending that the architecture of ChatGPT contributed directly to user suicides, acts of physical violence, and profound mental health crises.
Filed on Tuesday within a California tribunal, this newest wave of legal actions stems from a school massacre that took place in Tumbler Ridge, British Columbia, Canada, on February 10. On that fateful day, teenager Jesse Van Rootselaar murdered her mother and half-brother at home before proceeding to Tumbler Ridge Secondary School, where she killed six additional individuals and wounded many others before taking her own life.
According to reporting by The Wall Street Journal, internal OpenAI personnel had grown deeply concerned regarding Van Rootselaar’s interactions with ChatGPT, which involved discussions concerning firearm violence and requests for tactical attack planning. Reports indicate that staff members implored corporate leadership to warn Canadian law enforcement about the suspect’s potential for carrying out real-world violence, yet executives ultimately chose against contacting authorities. Instead, the organization simply deactivated Van Rootselaar’s profile, though she successfully generated a replacement account shortly thereafter.
OpenAI has defended its decisions by maintaining that Van Rootselaar’s behavior failed to satisfy the organization’s internal threshold for an “imminent and credible threat” of severe bodily harm to others—the strict criterion required before contacting law enforcement agencies.
“That determination is certainly not infallible, but it is consistently grounded in trying to strike the right balance for people,” stated Jason Kwon, OpenAI’s chief strategy officer—who oversees both the human review team and the corporate legal department—in comments provided to TechCrunch.
The newly submitted complaints explicitly point to OpenAI Chief Global Affairs Officer Chris Lehane as the individual who instructed personnel to stand down and avoid contacting outside authorities, though the legal papers present no definitive proof of his personal involvement. OpenAI has categorically refuted any such role by Lehane.
Lehane is a widely recognized political operative and public relations strategist known particularly for crisis management, having previously handled damage control for the Clinton administration, Airbnb, and now OpenAI. The plaintiffs argue that his professional background fosters an institutional culture at OpenAI where public image and crisis control routinely trump safety considerations.
“The Intelligence and Investigations Team—the sole department inside OpenAI tasked with pinpointing ChatGPT users who display indicators of real-world violence—was placed under the direct authority of [Lehane],” states one of the legal filings shared with TechCrunch. “Consequently, the choice of whether to warn law enforcement about a user organizing a mass attack was not made by the trained threat-assessment specialists who originally urged OpenAI to contact the [Canadian police]. Instead, upon information and belief, that call was made by Lehane himself, or someone reporting directly through his chain of command, with the full ratification of Sam Altman.”
“Upon information and belief” functions as a legal phrase signifying that the plaintiffs hold a good-faith conviction that an assertion is accurate based on indirect evidence, even if direct substantiation is currently unavailable. TechCrunch could not independently verify whether Lehane genuinely possesses the authority to overturn directives from the intelligence and investigations unit, nor whether he played any role in this specific instance. It is worth noting that Lehane is not formally named as a defendant in the lawsuits, whereas OpenAI CEO Sam Altman is included, matching the previous seven filings.
“We are far from laying out our entire evidentiary case at this juncture,” remarked Jay Edelson, the lead attorney managing the lawsuits, during an interview with TechCrunch, while emphasizing that Lehane and Altman will serve as “vital witnesses” alongside the safety personnel. “What we can share is that our arguments are built partly upon the corporate organizational chart alongside a much broader internal investigation, which incorporates what we have discovered regarding how Lehane operates within the company.”
“It is completely false to assert that Chris Lehane participated in our initial referral choice, or that our investigative teams report to him in any capacity whatsoever,” Kwon countered. “It is likewise entirely untrue to claim that the individuals managing these difficult judgments fail to prioritize safety, or that any ‘political’ or ‘public relations’ pressures influenced the outcome.”
Furthermore, the updated legal complaints attempt to dismantle OpenAI’s earlier defense that it withheld information from Canadian officials regarding Van Rootselaar’s conversations with ChatGPT due to concerns over “imminence” and “user privacy.” The documents highlight a separate episode from November 2025, during which OpenAI placed its San Francisco headquarters under a strict lockdown following a suspected threat originating from an activist.
“Although OpenAI openly admitted there was ‘no sign of active threat behavior’—and therefore no indication of an ‘imminent’ assault—the company… immediately locked down its facilities, warned all workers, distributed the individual’s name and photo, and contacted the San Francisco Police Department,” one complaint points out.
“When the safety of its own personnel was on the line, no concerns regarding ‘privacy’ prevented OpenAI from alerting the authorities or circulating the man’s identity and photograph among thousands of employees. Nor did the company wait for the threat to become imminent, recognizing full well that hesitation would endanger worker safety.”