Overview
- OpenAI removed the “Make this chat discoverable” checkbox on shared ChatGPT links after Fast Company revealed thousands of conversations in Google search results.
- CISO Dane Stuckey called the feature a “short-lived experiment” that risked accidental exposure and announced its removal on July 31.
- Indexed chats included sensitive personal and proprietary data such as mental health disclosures, legal confessions and corporate code.
- OpenAI is working with Google, Bing and other search engines to de-index affected links, though cached versions may remain accessible for some time.
- Meta AI continues to allow public chats in its Discover feed to be indexed by Google, underscoring divergent privacy defaults across AI platforms.