OpenAI debated calling police about suspected Canadian shooter’s chats

Summary

An 18-year-old, Jesse Van Rootselaar, who allegedly killed eight people in a mass shooting in Tumbler Ridge, Canada, reportedly used ChatGPT in ways that alarmed OpenAI staff, according to The Wall Street Journal.

Chats describing gun violence were flagged by tools that monitor misuse of the company’s LLM and the account was banned in June 2025. Staff debated whether to contact Canadian law enforcement but ultimately did not.

After the incident, OpenAI says it proactively reached out to the Royal Canadian Mounted Police with information about the individual’s use of ChatGPT and will continue supporting the investigation.

In practice

The report illustrates how abuse-monitoring systems can flag troubling content and trigger enforcement actions like bans without necessarily leading to a law-enforcement referral.

OpenAI says the activity did not meet its criteria for reporting to law enforcement before the attack, and that it contacted Canadian authorities after the tragedy.

The article also points to other warning signs in the suspect’s digital footprint, including creating a Roblox game that simulated a mall mass shooting and posting about guns on Reddit.

What we still don't know

The report does not spell out the exact criteria OpenAI uses to decide when to notify authorities, or what information can be shared in different scenarios.

It also does not provide the specific chat content that triggered the internal alerts and ban, or whether any other interventions were considered beyond account enforcement.

More detail is also missing on how signals from multiple platforms and prior police interactions might be weighed in risk assessments.

Why it matters

• Highlights trade-offs between safety monitoring, privacy, and timely intervention.
• Increases scrutiny on clear, consistent escalation policies for credible threats.
• Underscores the broader debate on AI systems and user safety in high-risk contexts.