OpenAI Considered Alerting Police Months Before Canadian Mass Shooting Suspect Acted

Months before Jesse Van Rootselaar allegedly carried out a mass shooting in the rural British Columbia community of Abbotsford, employees at OpenAI flagged troubling patterns in her interactions with ChatGPT, according to multiple reports. The AI company considered alerting Canadian law enforcement as early as eight months prior to the attack, but ultimately chose not to intervene, a decision that has since ignited a global conversation about the ethical obligations of tech companies in the face of potential real-world violence.

According to The Straits Times, internal communications at OpenAI revealed that Van Rootselaar’s use of the chatbot included repeated queries about firearms, tactical planning, and expressions of violent ideation. These interactions, detected through automated monitoring systems, triggered a high-priority alert within the company’s safety team. Employees reportedly urged leadership to contact authorities, citing the potential for imminent harm. However, legal and policy concerns—particularly around user privacy, jurisdictional boundaries, and the absence of a clear legal mandate—led to a decision not to escalate the matter.

As reported by MSNBC, the internal debate was intense. Some engineers and ethicists argued that OpenAI’s duty of care extended beyond digital boundaries, especially when users exhibited signs of planning mass violence. Others warned that preemptive reporting could set a dangerous precedent, normalizing surveillance of private conversations and chilling free expression on AI platforms. The company’s legal team reportedly consulted with external counsel, concluding that without a direct threat or identifiable target, no legal obligation existed to notify law enforcement under existing U.S. or Canadian statutes.

The tragedy unfolded in late 2025, when Van Rootselaar opened fire in a community center, killing seven people and injuring several others. In the aftermath, investigators discovered a digital trail linking her to ChatGPT, including prompts such as: "How can I maximize casualties in a public space?" and "What are the most effective weapons for a rural ambush?" These messages, according to court documents, were stored in her account and later retrieved by authorities during a forensic review.

OpenAI has since acknowledged the missed opportunity. In a statement to the press, the company said: "We are deeply saddened by the loss of life and are conducting a full internal review of our threat detection and escalation protocols. We recognize that our systems, while robust, are not infallible—and that ethical responsibility must evolve faster than regulation."

The case has prompted calls from lawmakers, mental health advocates, and AI safety researchers for new frameworks governing AI companies’ responsibilities. The Canadian government has launched an inquiry into whether tech firms should be legally required to report credible threats detected through their platforms. Meanwhile, the U.S. Senate has scheduled hearings on AI and public safety, with OpenAI’s CEO expected to testify.

Experts warn this incident may be a harbinger of a new class of crimes—where AI tools are weaponized not just for misinformation, but for operational planning. "We’re no longer just dealing with bots spreading lies," said Dr. Elena Torres, a cybersecurity ethicist at Stanford University. "We’re now confronting AI as a co-conspirator in real-world violence. The question isn’t whether platforms should monitor—but how, and with what accountability."

As communities mourn the victims in Abbotsford, the world watches to see whether OpenAI—and the broader AI industry—will transform this failure into a catalyst for systemic change, or if another tragedy will be required before meaningful reform takes hold.