在Anthropic向执法部门报告了一名妇女与克劳德的谈话后，治安官的副手在佛罗里达州逮捕了一名妇女。据称，她在聊天中表示将向李县治安官办公室“开枪”。WINK News于9月30日首次报道了这起逮捕事件。
Sheriff’s deputies have arrested a woman in Florida after Anthropic reported her Claude conversation to law enforcement. In the chat she allegedly said she would “shoot up” the Lee County Sheriff’s Office. WINK News first reported the arrest on 30 September.
根据逮捕报告，她于9月26日撰写了这份声明。在第二天的一条消息中，她说她有一把新枪。
According to the arrest report, she wrote the statement on 26 September. In a message the following day, she said she had got a new gun.
调查人员在报告中写道，Anthropic的平台监控关键短语和潜在威胁内容。由于声明的严重性，这些声明交给了一个人类审查团队。该团队向执法部门报告了他们。“从来没有真正的匿名”代表们将这名妇女拘留在家中，没有发生任何事件。治安官办公室的一名情报侦探随后接管了此案。警长Carmine Marceno告诉WINK新闻，她后来说她像“日记”一样使用人工智能。
Anthropic’s platform monitors for key phrases and potentially threatening content, investigators wrote in the report. Because of their severity, the statements went to a human review team. That team reported them to law enforcement. ‘Never truly anonymous’Deputies detained the woman at her home without incident. An intelligence detective from the sheriff’s office then took over the case. She later said she uses AI like a “diary”, Sheriff Carmine Marceno told WINK News.
“人工智能是一种强大的工具，就像任何技术一样，它也可能被滥用。当有人使用人工智能制造或促成威胁时，我们必须认真对待该威胁。用户需要明白，你永远不是真正的匿名者，”马塞诺在一份声明中说。
“Artificial intelligence is a powerful tool, and like any technology, it can be misused. When someone uses AI to make or facilitate a threat, we have to take that threat seriously. Users need to understand that you are never truly anonymous,” Marceno said in a statement.
据Tom ' s Hardware报道，法庭记录列出了9月30日根据佛罗里达州法规836.10提出的重罪指控。该法规涵盖大规模枪击或恐怖主义行为的书面或电子威胁。法院任命了一名公设辩护人，她的提审定于11月2日进行。
Court records list a felony charge filed on 30 September under Florida Statute 836.10, Tom’s Hardware reported. The statute covers written or electronic threats of a mass shooting or act of terrorism. The court appointed a public defender, and her arraignment is set for 2 November.
据Tom ' s Hardware称，Anthropic尚未对此案公开发表评论。
Anthropic has not commented publicly on the case, according to Tom’s Hardware.
Anthropic的政策称Anthropic的隐私政策自9月10日起生效。该公司表示，如果Anthropic真诚地相信披露对于“防止对任何人或财产造成严重伤害”是“合理必要的”，则可以与执法部门共享个人数据。
What Anthropic’s policies say Anthropic’s privacy policy has been in force since 10 September. It says Anthropic may share personal data with law enforcement where it believes in good faith that disclosure is “reasonably necessary” to “prevent serious harm to any person or to property”.
其针对政府请求的政策规定，仅在有效的法律程序下披露用户信息。例外情况是“可能导致迫在眉睫的身体伤害或死亡”的紧急情况。
Its policy on government requests says it discloses user information only under valid legal process. The exception is an emergency that “may result in imminent physical harm or death”.
引用地区新闻报道，Tom ' s Hardware的早期案件统计了自8月份以来警方收到的至少两次克劳德早期对话。8月11日，在圣安东尼奥，一名22岁男子询问“Human pic AI聊天”，询问附近一所小学发生的枪击事件。在联邦调查局的警报下，警方以恐怖威胁重罪指控逮捕了他。
Earlier cases Tom’s Hardware counts at least two earlier Claude conversations that reached police since August, citing regional news reports. In San Antonio, a 22-year-old man asked “an Anthropic AI chat” on 11 August about shooting at a nearby elementary school. Police, alerted by the FBI, arrested him on a felony terroristic threat charge.
8月14日，在旧金山，一名用户在克劳德聊天中威胁Anthropic首席执行官达里奥·阿莫迪（Dario Amedei）。他说他买了一架AR-15。Anthropic通知了警方，但警方没有逮捕或指控他。
On 14 August in San Francisco, a user threatened Anthropic chief executive Dario Amodei in a Claude chat. He said he had bought an AR-15. Anthropic notified police, who did not arrest or charge him.
OpenAI还报告了ChatGPT用户。据《棕榈滩邮报》报道，今年3月，一名佛罗里达州男子写信给ChatGPT，要求杀害他的前女友，OpenAI向联邦调查局举报了他。根据同一法规，他于八月承认了包括书面威胁杀人在内的指控，并被判处八年缓刑。
OpenAI has also reported ChatGPT users. In March, a Florida man wrote to ChatGPT about killing his ex-girlfriend, and OpenAI reported him to the FBI, the Palm Beach Post reported. He pleaded guilty in August to charges including written threats to kill, under the same statute, and was sentenced to eight years’ probation.
今年早些时候，萨姆·奥尔特曼（Sam Altman）在OpenAI没有向警方通报Tumbler Ridge枪手的账户后道歉。佛罗里达州总检察长于6月起诉了OpenAI，目前正在寻求法院针对该公司的命令。
Earlier this year, Sam Altman apologised after OpenAI did not alert police about the Tumbler Ridge shooter’s account. Florida’s attorney general, who sued OpenAI in June, is now seeking a court order against the company.