What Happened? The “Incident” in Summary
For a short duration, ChatGPT conversations were made discoverable by search engines in July 2025.
This feature was optional and required users to manually check a box labeled “Make this chat discoverable” when sharing a conversation. Eventually, this led to a widespread discussion regarding privacy concerns.
In response to privacy concerns and reports of nearly 4,500 conversations appearing in Google search results, OpenAI removed the feature on August 1, 2025..
What Went Wrong: Shared ChatGPT Conversations Got Indexed by Google
1. Publicly-shared links became searchable.
Users could click the “Share” button in ChatGPT and optionally select the “Make this link discoverable” checkbox. That made the conversations not just accessible via URL—but indexable by Google and other search engines.
2. Thousands of conversations appeared in search results.
Reporters uncovered over 4,500 indexed ChatGPT conversations, some containing deeply personal info like mental health struggles, trauma, or career details.
3. Google didn’t do this on its own—it merely indexed what was made public.
Since users made those URLs publicly accessible, Google crawled them like any other webpage. OpenAI and users alike were surprised by the outcome.
OpenAI’s Response: Feature Removed and Cleanup Underway (August 1st)
- The feature was rapidly shut down.
Within hours of media reports, OpenAI disabled the option to make chats discoverable. - Worked to remove indexed content.
OpenAI began purging these shared chat links from search engine results, though some cached versions may persist temporarily
So is Our Conversation with ChatGPT safe now?
Fortunately, together with the above mentioned “fix”, right now, your private chats are not indexed by search engines — meaning they aren’t visible on Google, Bing, or any public web search.
Which means they are kind of safe!
However, OpenAI may use some conversations to improve models if you’ve enabled training data sharing in your settings (and personal info is removed), but those aren’t public-facing or searchable.
Why Does This Matter?
- Privacy risk—even if unintentional.
Many users didn’t realize how easily their private ChatGPT conversations could become public—and visible to any search engine user. The checkbox default wasn’t misleading, but its design and wording were subtle enough to cause confusion.prompt.securityTechRadar - A reminder: treats AI chats like email or cloud docs.
Whatever you share can be saved, visible, or indexed—especially when using AI tools with public-sharing capabilities.
