OpenAI's Altman warns AI chats lack legal privacy protections

OpenAI CEO Sam Altman has warned that conversations users hold with AI chatbots are not protected by legal privilege, meaning deleted messages can still be retained on company servers for a month or longer and produced as evidence in court when ordered. Altman noted that ChatGPT is widely used for psychological support but said the company offers no legal guarantee for such use. Privacy experts have cautioned that mental and chronic health information shared with chatbots could be accessed by insurance companies for risk assessment purposes. Financial data, including location, bank card and account numbers, and salary details, could also be exposed, according to the warnings. Uploaded photographs may be used for model training or to generate fake content, and passwords and verification codes can be logged by the platforms. Software engineer Hizam al-Subaii told the outlet that AI conversations can be handed over to governments or courts upon request. No further details were immediately available.