Overview

• OpenAI removed its “Make this chat discoverable” toggle and has worked with major search engines to de-index previously shared conversations
• Investigations by Fast Company and others found thousands of indexed ChatGPT links revealing mental health disclosures, internal job evaluations, crime confessions and proprietary code
• Internet Archive snapshots and search engine caches continue to host exposed conversations even after their removal from live search results
• The feature’s unclear opt-in design exposed shortcomings in consent mechanisms and underscores the need for secure-by-design privacy defaults in AI chat services
• Meta confirmed that its AI app’s public Discover feed still permits Google indexing, highlighting ongoing privacy vulnerabilities across platforms