Recently, a privacy controversy erupted when thousands of ChatGPT conversations, shared via public links, began appearing in Google search results. This surprising discovery revealed that if users utilized ChatGPT’s “Share” feature—which generates a unique public link for any conversation—those chats could become indexed by search engines like Google. As a result, dialogues containing everything from personal stories to sensitive work discussions, and even resumes or legal identities, became accessible to anyone performing a simple search.

The incident highlights how digital privacy can be compromised unintentionally. When a ChatGPT chat is shared using a public link, anyone with access to that link can view the conversation. If that link is posted on public forums, websites, or passed around, Google’s bots may crawl and index it, making it available through search queries. Unfortunately, even after deleting a shared chat or disabling the share link, the indexed page might remain in Google’s cache for some time before disappearing.

Although OpenAI does not display the name of the original user with the chat, sensitive information included within the messages—such as names, contact details, or confidential business data—could still be exposed. This led to significant user backlash and privacy concerns.

In response to the outcry, OpenAI quickly withdrew the experimental feature that enabled these conversations to be “discoverable” by search engines. The company clarified that the setting to make chats readable by search engines was always optional and clearly labeled. Still, many users were unaware of the risks involved, and OpenAI ultimately decided to prioritize user privacy and remove the web discoverability option altogether.

This episode serves as a critical reminder: when using AI tools, always consider the visibility and permanence of whatever you choose to share. Avoid including sensitive or personally identifiable information in any content you generate and be cautious with features enabling public sharing. Treat AI conversations with the same discretion as you would public posts or cloud-stored documents.