A 29-year-old Sydney man has been charged after allegedly threatening to carry out a mass shooting at a childcare centre, with his conversations with an artificial intelligence chatbot reportedly flagged to authorities.
David Lan Thackray, from Hurstville in southern Sydney, was arrested on 18 October and charged with using a carriage service to menace, harass or offend, according to media reports.
According to a police fact sheet tendered to the court, Thackray allegedly told Claude, an AI chatbot developed by Anthropic, that he intended to shoot children at a nearby childcare centre.
The Sydney Morning Heraldreported that he allegedly threatened
The Sydney Morning Heraldreported that he allegedly threatened to go to a local Westfield shopping centre and “shoot up the local daycare”.
It is further reported that the court also heard allegations that he discussed targeting a doctor’s office and referenced Naveed Akram, who has been charged over the December 2025 Bondi Beach mass shooting that killed 15 people.
Prosecutors argued against releasing Thackray on bail, warning of potentially catastrophic consequences if he acted on the alleged threats. They told the court his stated intentions were not vague and raised concerns about what he might do if released.
It is reported that defence lawyer Danilo Rajkovic
However, it is reported that defence lawyer Danilo Rajkovic argued that there was little substance behind his client’s words, describing the chatbot exchanges as a stress-induced outburst. He said Thackray had initially been using the chatbot in an attempt to address an ear infection.
The defence also noted that Thackray was studying international relations, had no criminal record and had lived in Australia for more than 17 years.
The case has drawn attention to how AI companies respond when users discuss potential acts of violence and under what circumstances information may be escalated to law enforcement.
Claude is designed to refuse requests involving
Claude is designed to refuse requests involving the planning or encouragement of violence. Anthropic’s policies also provide for user data to be shared with law enforcement in certain emergency circumstances.
Media reports said Thackray’s conversations were flagged through Interpol before his arrest. The precise circumstances of the referral and the steps taken to assess the alleged threat have not been fully established in the material available.
As perChannel 10, Thackray has pleaded not guilty. He was remanded in custody for a mental health assessment after appearing at Sutherland Local Court on 19 October.
No-nonsense journalism. No paywalls. Whether you’re in Australia, the UK, Canada, the USA, or India, you can support The Australia Today by taking a paid subscriptioness journalism alive
