Private conversations with ChatGPT have been turning up in Google search results, raising alarm over how easily personal information can slip into the public domain when AI tools are used for sensitive discussions.
The problem surfaced when OpenAI tested a “discoverable” setting that let users deliberately share chats online. Anyone who ticked the box marked “make this chat discoverable” was told it would “be shown in web searches.”
While the feature stripped names from the transcripts, it did not remove the deeply personal nature of many exchanges, from confessions about harassment to candid fears and therapy-like conversations.
Journalist Luiza Jarovsky, who spotted the issue, posted on X that she had found examples of these personal exchanges appearing in Google’s index. Her findings showed how a simple misclick or hasty decision could leave private thoughts permanently searchable.
On Thursday, OpenAI’s chief information security officer, Dane Stuckey acknowledged the situation publicly. “We just removed a feature from @ChatGPTapp that allowed users to make their conversations discoverable by search engines, such as Google,” he wrote.
Stuckey said the tool had been “a short-lived experiment to help people discover useful conversations” but admitted it created “too many opportunities for folks to accidentally share things they didn’t intend to.”
He added that the company was working to have existing indexed content removed from search engines, with the change set to reach all accounts by Friday morning.
Many people unfortunately turn to ChatGPT for emotional support, personal problem-solving, or to share experiences they might never voice aloud to another human. Yet once shared through an online platform, even accidentally, those same conversations can become public artifacts.
