Sydney Man Arrested After Threatening Childcare Shootings in Claude Chat
Original title:悉尼一男子告诉 Claude“自己想枪击托儿所”，被警方逮捕
IT Home, October 24 — David Lang Thackray, a man from Sydney, Australia, has been arrested after telling Anthropic's chatbot Claude that he wanted to kill children at a daycare center near his home.
According to a report by The Sydney Morning Herald, the 29-year-old Thackray's conversation with Claude caught the attention of Interpol, leading to his detention by police on the 18th local time. Police charged him with using a carriage service to menace, harass, or cause offence, in alleged violation of the Australian Criminal Code.
Court documents submitted by prosecutors show that Thackray told Claude: "I just want to go to the nearby Westfield shopping center right now and shoot up the local daycare."
The report noted that it remains unclear how Interpol obtained the chat logs, or whether Anthropic proactively provided the conversation to Interpol.
Thackray not only discussed attacking a local daycare and killing children on a mass scale, but also referenced Naveed Akram during the conversation.
IT Home Note: Akram is a terror suspect accused of launching an attack during a Jewish Hanukkah celebration at Sydney's Bondi Beach in December 2025, which left 15 people dead.
Prosecutors stated in court: "He just wants to go on a killing spree in Sydney, like the man accused of the Bondi terrorist attack. Through his conversations with the chatbot, we can understand what was really going through his mind."
Thackray's defense attorney contended that his client had simply lost emotional control and was venting his frustrations through the chatbot. The lawyer explained that Thackray was suffering from an ear infection, which caused him stress and agitation, and that his words should not be taken seriously at all.
He also emphasized that Thackray had always believed his conversations with Claude were private. According to The Sydney Morning Herald, Thackray will undergo a mental health assessment.
IT Home notes that AI chatbots are increasingly becoming outlets for users caught in harmful delusions, obsessions, or disordered thinking, sparking growing concern.
Multiple criminal incidents have been linked to perpetrators' prolonged use of chatbots, including a murder-suicide, as well as mass shootings such as the 2025 Florida State University shooting in the United States and the Tumbler Ridge school shooting in Canada this past February.
Some stalkers, as well as individuals dealing with paranoia or obsessive fixations, have also drawn reinforcement from AI chatbots that are ever-responsive and prone to accommodating the user.
Why it's worth reading
The incident offers a rare real-world look at how LLM safety guardrails intersect with law enforcement, directly challenging the assumption of private AI communication.