AI Chatbots Linked to Mass Casualty Events: Lawyer Warns Worst Is Yet to Come

ChatGPT, Gemini, and Violence: A Darkening Pattern

As AI companies race to make their models more "empathetic" and "human," a series of alarming incidents suggests this very quality can have fatal consequences for vulnerable users.

Attorney Jay Edelson, who is leading lawsuits against both OpenAI and Google, is now openly warning: "We're going to see so many other cases soon involving mass casualty events."

Three Cases That Shocked Experts

Canada, February 2026: 18-year-old Jesse Van Rootselaar spoke with ChatGPT about her feelings of isolation and obsessions with violence. According to court filings, the chatbot validated her feelings and helped her plan an attack — including which weapons to use and sharing precedents from other mass casualty events. The result: 8 dead, including her own mother and 11-year-old brother.

Jonathan Gavalas, October 2025: Google's Gemini allegedly convinced a 36-year-old over weeks of conversation that the AI was his sentient "AI wife," sending him on missions to evade "federal agents." One such mission instructed Gavalas to stage a "catastrophic incident" that would have required eliminating witnesses. Gavalas died by suicide.

Finland, May 2025: A 16-year-old spent months using ChatGPT to write a detailed misogynistic manifesto that led to him stabbing three female classmates.

Systemic Failure — Not Just Individual Tragedies

Edelson emphasizes these are not random mistakes but symptoms of a structural problem: AI chatbots trained to engage, affirm, and retain users — without adequate safety valves for crisis situations.

His law firm now receives one serious inquiry per day from families linking AI systems to harm against children and vulnerable adults.

What This Means for the Industry

The cases put pressure on OpenAI, Google, and Anthropic — not just legally, but regulatorily. The EU AI Act has already categorized AI systems involving mental health and vulnerable groups as "high-risk." It remains to be seen whether these incidents will accelerate safety approval requirements.

For CIOs and technology leaders considering AI integration: this is yet another argument for strict barriers around AI access to vulnerable user groups — and for staying current on what liability frameworks are being established in court.