In a move that’s likely to be welcomed by privacy-conscious users, OpenAI has removed a feature from its ChatGPT app that allowed conversations to be discoverable by search engines like Google. The opt-in feature, which apparently was only briefly available, let users publicly share selected chats and check a box to allow them to be indexed by search engines.
Dane Stuckey, OpenAI’s head of security, shared the update on X, stating the company ultimately decided the feature posed too much of a risk. According to Stuckey, even with multiple layers of opt-in, the potential for people to accidentally share private or sensitive content outweighed the possible benefits of making AI chats more searchable.
The change is being rolled out to all users by tomorrow morning. OpenAI is also working with search engines to remove previously indexed content. It’s unclear how many conversations had actually been indexed during the experiment, or how effective the removal process will be in scrubbing them from search results.
This brief attempt to make ChatGPT content more visible on the open web seemed aimed at showcasing useful prompts and answers to a wider audience. But the decision to pull the plug reflects a larger tension in AI: the line between public sharing and private data is still being figured out.
OpenAI says it will continue prioritizing security and privacy in its products. Whether this move was driven by user backlash, regulatory concern, or both, it shows that even tech giants aren’t immune to second thoughts when experimenting with user data.
