WhatsApp’s hidden instructions for Meta chatbot about 'WHAT NOT TO SAY' to users leaked

AI Summary

WhatsApp accidentally exposed internal guidance for its AI chatbot across about 80 help and FAQ pages. Meta said a bug caused the disclosure and has been fixed; the instructions included a prohibition on giving legal advice or interpretations.

WhatsApp unintentionally exposed internal instructions for its AI chatbot within around 80 help and FAQ pages. These hidden instructions guided the chatbot on handling various support queries, including account deletion and legal issues. Meta clarified that the exposure was due to a bug which has since been fixed. Notably, the chatbot was instructed not to provide any legal advice or interpretations.