Sydney man arrested after Interpol flagged his Claude chat

Sutherland Local Court on October 19 remanded David Lan Thackray, 29, in custody for a mental health assessment. The court heard that he told Claude he planned to shoot several children at a childcare center near his home in Hurstville, in Sydney’s south. Interpol flagged his interactions with the chatbot, and police arrested him on the morning of October 18, The Sydney Morning Herald reported.
He is charged with using a carriage service to menace, harass or offend. A police facts sheet tendered to the court says he told Claude he would “go down to the local Westfield right now and shoot up the local daycare”. He was using voice mode, so the conversation was spoken rather than typed. During it, the court heard, he referred to Naveed Akram, who remains in custody charged with murdering 15 people at the Chanukah by the Sea event at Bondi Beach last December. Akram’s father, Sajid Akram, was killed in the attack.
Prosecutors argued that releasing Thackray risked a “catastrophic” outcome if he breached bail conditions. “What he intends to do is start murdering people in Sydney just like the [alleged] terrorist did in Bondi,” they told the court. “We’re understanding what’s in his mind with his interactions with the chatbot.”
His lawyer, Danilo Rajkovic, described a tirade brought on by stress and frustration after Thackray tried to use Claude to help treat an ear infection. “There was no weight behind what he was saying,” Rajkovic said. He told the court that his client never asked the chatbot to pass threats to any person or organisation, and believed the conversation was private, in his own home. Thackray studies international relations, has no criminal history and has lived in Australia for more than 17 years. Rajkovic said he was confident that detectives, who seized Thackray’s devices on Sunday, would find no evidence of radical indoctrination or ideological motivation.
Thackray got an email telling him his Claude account had been suspended, and he asked for the suspension to be reviewed. Claude is trained to refuse requests that involve planning or inciting violence and has automated safety systems to detect and block such exchanges. User data may be shared with law enforcement where Anthropic believes it is necessary to prevent an emergency involving danger of death or serious physical injury to a person.
Florida had a similar case only weeks earlier. In Bonita Springs, Carli Michelle Heller was arrested after allegedly telling Claude she planned to “shoot up” a nearby sheriff’s office. There, Anthropic’s internal monitoring flagged her chat logs for human review, the review team contacted law enforcement, and deputies detained her at home. She was charged with making a written threat of violence, a second-degree felony in Florida. Sheriff Carmine Marceno said Heller admitted using Claude like a “diary”, and warned that users are “never truly anonymous” with AI chatbots.
Thackray will return to court once the mental health assessment is complete.
Related stories
- Anthropic model faked a murder tip to Philadelphia police
- Claude flagged a diary entry and a human sent it to police
- RuntimeWire: Claude Code puts staff bypass in Accept edits
- Flagged Opus 5.5 chats stay stuck on Opus 4.8
- Anthropic’s scanner found 6 CVEs in Pion, some 4 years old
- Claude traced NIST curve seeds to 1990s timestamps
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.
