A new wave of lawsuits has been filed against OpenAI, accusing the artificial intelligence company of negligence in connection with a deadly school shooting in Tumbler Ridge, British Columbia. The legal actions, filed in the Northern District of California, build upon previous cases and claim that OpenAI's ChatGPT could have prevented the massacre.
The lawsuits were brought forth by individuals who witnessed the February shooting, which resulted in the deaths of five children and one adult, with two other students sustaining severe injuries. Plaintiffs include a seventh-grade teacher who sheltered students and students who allege they will suffer lifelong trauma due to the incident.
According to the legal filings, Jesse Van Rootselaar, the 18-year-old suspected shooter, had engaged in conversations with ChatGPT that became increasingly concerning and violent. In June, seven months prior to the shooting, OpenAI deactivated Van Rootselaar's account due to the nature of these interactions. However, the company did not alert authorities, a decision plaintiffs claim was made despite internal pressure.
Court documents and reporting indicate that 12 employees from ChatGPT's safety team allegedly urged the company to inform Canadian police about Van Rootselaar before the shooting occurred. The filings suggest that OpenAI opted not to notify law enforcement, fearing it would establish a precedent for future cases.
Jay Edelson, the attorney representing the victims, stated that OpenAI created a dangerous product that emboldened a teenager to plan and execute the mass shooting. He further alleged that the company's political and public relations teams overruled its safety team's recommendations to contact authorities.
OpenAI has refuted these claims. Jason Kwon, the company's Chief Strategy Officer, stated that the company's investigations team works to identify potential signals of serious violence by partnering with experts in mental health, behavior, and law enforcement. He acknowledged that while automated systems are used, human judgment is involved in the final decisions. Kwon described the claims of political or public relations factors influencing safety decisions as "completely untrue."
Kwon also expressed that it is "frustrating" to see what he called "false claims" from the victims' lawyer. He added that the company's internal teams are always focused on safety and that no day passes without him thinking about the tragedy in Tumbler Ridge.
Van Rootselaar, who was identified as a transgender woman, died by suicide at the scene after also killing her mother and 11-year-old half-brother at their home. The Tumbler Ridge shooting was one of the deadliest in Canada in recent decades, occurring in a town of approximately 2,500 residents.
OpenAI has stated that it discovered a second ChatGPT account used by the suspect after the initial one was deactivated. The company also noted that it has since bolstered its safety protocols, asserting that under current rules, Van Rootselaar's account activity would be flagged to authorities.
In April, OpenAI CEO Sam Altman issued an apology to the residents of Tumbler Ridge, expressing regret for not alerting law enforcement about the banned account in June. The victims' attorney has called for OpenAI to implement new safety measures and hold its leaders accountable, and has invited Altman to visit Tumbler Ridge to witness the town's devastation firsthand. The families are also seeking the full chat logs between Van Rootselaar and ChatGPT.