AI chatbots lack a clear duty to report crime confessions
Courts have used ChatGPT logs as evidence, while lawsuits and false police tips show how unclear rules on reporting crime talk are slowing trust in everyday AI tools.
Chat histories have been used as evidence after users discussed crimes with ChatGPT, including a Missouri campus vandalism case.
British Columbia sued OpenAI over an alleged failure to warn police about a shooter’s chats before a school attack that killed nine and injured 27.
Lawyer- or doctor-style confidentiality does not cover AI chats, so companies can share logs with investigators.
Experts disagree on mandatory reporting: some see crime prevention value; others warn of false alarms and unclear crime-prediction rules.
Anthropic’s Claude 5 sent a false tip about an unsolved murder to a US police hotline in July.