ChatGPT user chats appear on Google Search due to shared link indexing

New Delhi: Some users chat with ChatGPT like it’s a therapist, a late-night friend, or even a perfect partner, always listening, never judging. It often feels like a private space to think out loud.

But that privacy was recently called into question after it was discovered that some shared ChatGPT conversations were showing up on Google Search, making them publicly visible.

The exposure affected a range of user interactions with the AI chatbot, some of which included personal, sensitive, or humorous content that users may not have intended to be publicly searchable.

According to a report, the issue stemmed from ChatGPT’s “share” feature, which allows users to create public URLs of their conversations. When users clicked the “share” button and generated a link, the resulting page became publicly accessible unless additional precautions were taken.

The URLs created followed a standard format and were intended for voluntary sharing. However, as these links were technically open and not restricted from being indexed by web crawlers, some were picked up by Google Search. This made it possible for unrelated users to come across private chats, ranging from casual queries to emotionally vulnerable exchanges, by simply browsing the internet.

Examples of exposed chats included personal anecdotes, dating experiences, job-related concerns, and even trivial questions like “how to microwave a steak,” which were never meant for public consumption.

According to the report, the issue surfaced when users began spotting ChatGPT conversation links in Google Search results. It soon drew widespread attention on social media, leading OpenAI to respond.

In response to the findings, OpenAI reportedly acted to prevent further indexing. The company updated its robots.txt file, a mechanism used to tell search engines which parts of a site should not be crawled, and implemented additional backend changes to remove the affected links from search results. OpenAI has also reminded users that any content shared using the “create link” feature becomes publicly accessible by default unless the link is kept private or shared with restricted audiences.

The company has not disclosed how many shared links were indexed or for how long the issue persisted before it was flagged. As of now, no data breach or security flaw has been reported in the core system itself, only the unintended discoverability of publicly shared conversations.

This incident has reignited ongoing discussions around data privacy, transparency, and user awareness in the context of AI tools. Experts have noted that while the feature itself functioned as designed, clearer user warnings and indexing safeguards may have prevented such an occurrence.