OpenAI’s ChatGPT was reportedly utilized in unsettling ways by an 18-year-old who purportedly slew eight individuals during a deadly shooting spree in Tumbler Ridge, Canada, causing significant concern among the company’s personnel.
Systems designed to supervise the firm’s LLM for improper application identified and flagged Jesse Van Rootselaar’s dialogues depicting weapon-related aggression, leading to her being banned in June 2025.
According to the Wall Street Journal, personnel within the organization deliberated whether to alert Canadian police authorities concerning the conduct but ultimately refrained. An OpenAI representative stated that Van Rootselaar’s actions failed to satisfy the benchmarks for notification to legal authorities; however, the company did contact Canadian officials following the event.
Van Rootselaar’s online presence encompassed more than just unsettling ChatGPT transcripts. She evidently developed a game on Roblox, a virtual environment widely used by minors, which mimicked a deadly gun attack within a shopping center. Additionally, she shared content regarding firearms on Reddit.
Area law enforcement was also familiar with Van Rootselaar’s volatile state, having been summoned to her household subsequent to her igniting a blaze while intoxicated by unidentified substances.
Large language model conversational agents, crafted by OpenAI and its rival firms, have been charged with instigating psychological crises in users who become detached from actuality while engaging with artificial intelligences. Numerous legal actions have been initiated, referencing dialogue records that incite individuals to self-harm or proffer support for that outcome.
Should you be experiencing an emergency or contemplating self-harm, please dial or message 988 to contact the 988 Suicide and Crisis Support Line.
Techcrunch event
Boston, MA
|
June 9, 2026
{content}
Source: {feed_title}