Google indexing exposed: Is it safe to use ChatGPT? 7 things to know

For most purposes, using ChatGPT is deemed "generally safe".

However, the recent Google indexing issue came as a shock for end users. It turns out your ChatGPT therapy chats aren't so private after all, Sam Altman himself warned.

It's just the tip of the iceberg. It emerged recently that some ChatGPT conversations were being indexed by Google and other search engines, Tech Crunch reported.

This highlights potential privacy risks.

It’s not clear whether or not users were made aware of this from the start. 

OpenAI has addressed this specific problem. Still, users must exercise caution, especially when sharing sensitive information. 

Following are 7 things you need to know about the Google indexing of ChatGPT conversations:

From late July to early August 2025, shared ChatGPT conversations with "/share" URLs were indexed by Google and other search engines if users enabled the “discoverable” option. This allowed public access to chats via search queries, exposing potentially sensitive content.

OpenAI’s "Share" feature allowed users to generate links for conversations, with an option to make them discoverable. These public links were crawled and indexed by search engines, as they lacked restrictions like noindex tags.

The tech mag explained that the indexed chats included sensitive topics (e.g., mental health, resumes), raising concerns about user privacy. Users may not have understood that “discoverable” meant searchable by anyone online.

On July 31, 2025, OpenAI disabled the discoverability feature for shared chats and began working to remove indexed links from search engines, acknowledging the privacy oversight.

Estimates on X suggested up to 70,000 conversations were indexed, though this figure is unverified. A Google site search for “site:chatgpt.com/share” could reveal these chats before the fix.

Users had to actively share conversations and enable discoverability for indexing to occur.

However, the unclear interface meant many didn’t realise the implications, highlighting the need for better privacy controls.

The incident underscores challenges in the AI industry, including balancing innovation with privacy, managing AI-generated content in search ecosystems, and avoiding regulatory scrutiny. It also strains trust in OpenAI and raises questions about search engine indexing policies.

Indexing is the process by which search engines like Google collect, analyse, and store web content to make it retrievable in search results.  When a webpage (e.g., a ChatGPT shared conversation with a "/share" URL) is publicly accessible, search engine “crawlers” may index it, meaning the content is added to the search engine’s database.

Once indexed, the webpage can appear in search results when users query relevant keywords, unless the site’s owner uses tools like robots.txt or noindex tags to block crawling. 

In the ChatGPT case, shared conversations were indexed because they were made discoverable, exposing them to public searches until OpenAI intervened.