OpenAI banned shooter’s ChatGPT account in June 2025
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI identified and banned a ChatGPT account belonging to Van Rootselaar in June 2025 under its abuse‑detection and enforcement program that flags accounts used to further violence.
- OpenAI did not notify authorities at the time because the suspect’s activity did not meet its internal threshold for a credible or imminent plan to cause serious physical harm.
- OpenAI later contacted Canadian police after the February 12 shooting, offering information on Van Rootselaar and pledging continued support for the investigation.
- Van Rootselaar is suspected of killing eight people on 12 February 2025 at Tumbler Ridge Secondary School and injuring 27 others before taking his own life.
- Wall Street Journal reported that about a dozen OpenAI staffers debated whether to act on Van Rootselaar’s posts, with some urging leaders to alert authorities.
- OpenAI says it trains ChatGPT to refuse illegal‑activity requests and to discourage imminent real‑world harm, and it is reviewing its referral criteria after the case.
Why it matters: The decision not to notify police before the attack highlights a gap in OpenAI’s risk‑assessment threshold, potentially leaving law‑enforcement agencies without early warnings while the company maintains its policy of limiting alerts to imminent threats. It also puts pressure on OpenAI to refine its abuse‑detection criteria, and victims’ families and the public may question the adequacy of AI platform oversight.
