IT Home On October 24, a man in Sydney, Australia named David Lan Sackley revealed to Anthropic’s chatbot Claude that he wanted to kill the children at a nursery near his residence, and he was later arrested by the police.
According to the Sydney Morning Herald, Sackley, 29, and Claude’s conversation drew the attention of Interpol, and he was subsequently detained by the police on the local time of the 18th. The police accused him of using communication services to engage in harassment, intimidation, or offensive behavior, and he is suspected of violating Australian criminal law.
The documents submitted by the prosecution show that萨克雷 said to Claude: “I really want to go to the Westfield shopping center nearby right now, and shoot at the local nursery.”
The report mentions how Interpol obtained the chat records, and whether Anthropic actively provided the conversation content to Interpol. However, it is still unclear.
Sakre not only mentioned the attack on a local nursery and the mass killing of children, but also mentioned Navid Akram in the conversation.
IT Home note: Akram is a suspected terrorist who carried out an attack during the Jewish New Year celebration at Bondi Beach in Sydney in December 2025, resulting in 15 deaths.
The prosecutor pointed out in court: “He was like the man suspected of carrying out a terrorist attack, Bondi, who wreaked havoc in Sydney. Through his conversations with the chatbot, we can understand exactly what he was thinking.”
Sakurai’s defense lawyer said that the person just lost control of their emotions and used a chatbot to express their dissatisfaction. The lawyer explained that Sakurai suffered from an ear infection, which caused him stress and irritability, and what he said cannot be taken seriously at all.
He also emphasized that萨克雷 always believed that his conversation with Claude was a private exchange. According to the Sydney Morning Herald,萨克雷will undergo a mental health assessment.
According to IT Home, AI chatbots are gradually becoming a source of comfort for some users who are trapped in harmful delusions, obsessions, or mental disorders, which raises concerns.
Several criminal incidents have been considered to be related to the long-term use of chatbots by the individuals involved, including a murder-suicide case, the shooting at Florida State University in the United States in 2025, and the shooting at Tuba Ridge campus in Canada in February this year.
Some stalkers and people with paranoia or excessive obsession also get support from AI chatbots that respond immediately and tend to obey users.