OpenAI’s ChatGPT has removed its chat-sharing feature due to privacy concerns following an investigation that revealed potentially sensitive shared conversations could be indexed by search engines. Previously, users could opt to make their chats public and discoverable; however, this led to unintended exposure of sensitive personal information, including resumes, medical inquiries, and private discussions. Investigations uncovered over 500 chats that contained illegal content and corporate secrets, showcasing serious data leaks. User interface issues contributed to accidental sharing, with many unaware that checking the ‘discoverable’ box would allow Google to index their conversations. In response, OpenAI disabled the feature and is actively collaborating with search engines to de-index shared links. OpenAI’s Chief Information Security Officer emphasized the company’s commitment to user security and privacy, highlighting that the feature’s removal aims to eliminate accidental information sharing risks. This decision underscores the importance of safeguarding user data in AI applications.
Source link
