ChatGPT's new Locked Chats feature has a major privacy catch (www.techradar.com)

OpenAI has introduced a new "Locked Chats" feature for ChatGPT that enhances user privacy by allowing users to secure specific conversations behind a four-digit PIN or biometric authentication, making them inaccessible during casual browsing. Users can easily lock a chat through the menu, after which the conversations are hidden from regular history and search results. However, if users forget their PIN and need to reset it, all locked chats will be permanently deleted, raising concerns about the need for careful management of this feature. While Locked Chats offer an appealing layer of privacy, they come with significant limitations. Locked conversations may still be visible on previously opened devices and do not prevent other details from leaking into ChatGPT's memory if interactions occur before the chat is locked. Previously shared links to locked conversations remain accessible unless deleted, and sensitive information can be inadvertently remembered by the AI if not locked beforehand. As such, while the feature provides a measure of confidentiality, users must take extra precautions to ensure their privacy is fully protected, highlighting the ongoing challenges in managing conversational data in AI systems.