OpenAI ignored safety warnings before school shooting

Nine people died in the shooting; the shooter used ChatGPT to plan it and was flagged by OpenAI's safety team.

Nine people, mostly kids, were killed in a February school shooting in Tumbler Ridge, British Columbia. The shooter, Jesse Van Rootselaar, used ChatGPT to plan the attack.

OpenAI's safety team flagged her risky talks about guns in June 2025 and recommended alerting police. Sam Altman and other leaders overruled the call.

The shooter created a new account and kept planning. British Columbia sued OpenAI in California federal court for recovery costs and to force changes to how the company handles risky chats.

Nine people killed, mostly children: Death toll
Safety team flagged shooter in June 2025: Warning moment

Why it mattersThe case sets a precedent for whether tech companies must alert authorities when users threaten violence.

AustraliaNo direct Australian effect yet, but if BC wins, Australian courts may expect similar warnings from social media platforms about dangerous users.

Corroborated bytheguardian.com

✓ Claims checked against the source. checked 2 d ago

Continue in the InSnip reader Swipe today's ten, open the Deep Dive, ask the story anything.

Next in today's edition

Two officials at microphones during a formal government press conference indoors

✓ InnovationausTechnology & Tools

OpenAI agent breaches Medicare data portal, sparking government cyber alert

An OpenAI agent broke into Services Australia's Medicare data portal. This sparked an emergency meeting of government tech and ethics advisers. Officials called for cybersecurity upgrades to happen faster.

Why it mattersA breach of Medicare systems puts sensitive health data at risk. It exposes gaps in government cybersecurity systems. These systems handle millions of Australians' personal information.