Precrime: Months Before Massacre, OpenAI Worried About Canada’s Trans Mass Killer

Months before a Canadian man in a dress went on a Feb 10 rampage, killing his mother and half-brother at home before slaughtering five students and an education assistant at a secondary school where he was formerly a student, employees at OpenAI were deeply troubled by his interactions with the firm’s ChatGPT AI chatbot.

As first reported by the Wall Street Journal, Jesse Van Rootselaar’s ChatGPT activity was flagged by the company’s automated review system. When employees took a look at what he’d been up to over a several-day period in June 2025, they were alarmed. About a dozen of them debated what they should do.

Some were convinced Van Rootselaar’s descriptions of gun-violence scenarios signaled a substantial risk of real-world bloodshed, and implored their supervisors to notify police, according to the Journal’s unnamed sources. They opted against doing so, and a spokeswoman now says they’d concluded Van Rootselaar’s posts didn’t cross the threshold of posing a credible and imminent risk of serious harm. Instead, the company decided only to ban his account.

About seven months after his disturbing series of interactions with ChatGPT, police say he killed 8 people and injured 25 more before killing himself in the school he’d attended earlier. Van Rootselaar’s social media and YouTube accounts contained transgender symbolism as well as the online name “JessJessUwU” (a meme phrase that people may recognize from the bullet casings tied to the gay suspect charged in the assassination of Charlie Kirk).

Only after the bloody horror unfolded did OpenAI contact the Royal Canadian Mounted Police. The situation highlights the difficult position social media and AI platforms are in, as they struggle with balancing conflicting goals: protecting their users’ privacy and avoiding unnecessary interactions with police, versus preventing crimes up to and including mass murder. The Journal didn’t report specifics about Van Rootselaar’s ChatGPT interactions.

ChatGPT wasn’t the only online resource where he evidenced a potential for violence: He’d also used Roblox to build a game centered on carrying out a mass shooting at a shopping mall. Online activity aside, Van Rootselaar was already on the radar of local police, who made multiple visits to his home in response to mental health episodes, and even temporarily removing firearms from the property. An RCMP official said that, on multiple occasions, he was “apprehended for assessment and follow-up.”

Police say Van Rootselaar gender-transitioned about six years ago. Given he was 18 when he exploded into violence, that translates into the very young age of about 12. There’s no indication that the “transition” went beyond “identity” and clothing, and into the realm of hormones and other body-transforming measures. Online, he bemoaned the fact that his six-foot frame would render impossible his aspiration to be a “petite” woman.

On Reddit, Van Rootselaar often posted about his use of prescription and other drugs, and curiosity about 5-MeO-DMT, a hallucinogen nicknamed “toad venom.” He said he’d been diagnosed with ADHD, obsessive compulsive disorder, major depressive disorder and autism spectrum disorder, and was taking “Setraline 380mg (SSRI). I on rare occasion take 2mg of Risperidone for sleep purposes (anti-psychotic.)” Setraline is the generic version of Zoloft. He dropped out of school about four years ago.

Elsewhere on Reddit, in a post about his “right to be myself” and his “right to Hormone Replacement Therapy,” Van Rootselaar noted that at least other people support his “bare minimum…right to bear arms,” adding, “I’m a 15 year old trans person, transitioning from Male to Female. I ‘own’ 7 firearms, it’s cool.” 

What’s definitely not cool: nudging 12-year-olds down the gender-transitioning path.