Meta has patched a critical security flaw in its AI chatbot that allowed users to access private prompts and AI-generated responses of others. Discovered by researcher Sandeep Hodkasia, the bug was ...