Forgot password
Enter the email address you used when you joined and we'll send you instructions to reset your password.
If you used Apple or Google to create your account, this process will create a password for your existing account.
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Reset password instructions sent. If you have an account with us, you will receive an email within a few minutes.
Something went wrong. Try again or contact support if the problem persists.
Image by RyanDonegan, CC BY 2.0.

Claude AI chats accidentally exposed in Google search, revealing users’ most bizarre requests

A surprising discovery involving Anthropic’s Claude AI chatbot left users wondering how private their shared conversations really were. Hundreds of publicly shared Claude chats briefly appeared in Google search results, allowing anyone to view conversations that were never meant to reach a wider audience.

Recommended Videos

The exposed chats were not the result of a security breach or a direct hack. Instead, the issue came from Claude’s sharing feature, which allows users to create public links for individual conversations. Many users apparently did not realize that these shared pages could be indexed by search engines and found through specific searches.

According to Dexerto, the issue was first noticed after a Reddit user discovered that certain Google searches were displaying a large number of shared Claude conversations. The discovery quickly led users to explore the unexpected collection of AI interactions.

The weirdest part was what people were actually asking Claude

Once the conversations became searchable, users found a strange mix of everyday questions and unusual personal requests. Some chats involved normal coding assistance, while others revealed deeply personal or bizarre experiments with AI.

One of the most unusual examples involved a person asking Claude for help becoming a nine-tailed fox from Asian mythology. The chatbot responded with meditation ideas, visualization exercises, and roleplay suggestions to help the user embody the mythical creature.

Other exposed conversations included a lawyer asking if they needed to report a possible ethics violation. Another user was working on a crypto wallet project and accidentally revealed passwords during the process. A different chat involved a conspiracy theory claiming birds were secretly professional rappers behind Jay-Z’s music.

Users on the Reddit discussion that uncovered the issue also pointed out that this was not the first time AI conversations had appeared in search results. A similar incident was reported by Forbes in September 2025, when hundreds of shared ChatGPT transcripts were reportedly indexed by Google.

The Reddit thread also revealed that many exposed conversations contained adult content alongside cybersecurity-related discussions. Some users reported that the fix initially worked only for Google, while similar searches continued showing results on other search engines like Bing and Brave for a short period.

Google has since removed the indexed Claude conversations from its search results. The search terms that revealed the chats no longer appear to show the exposed pages.


Attack of the Fanboy is supported by our audience. When you purchase through links on our site, we may earn a small affiliate commission. Learn more about our Affiliate Policy
Author
Image of Santosh Kumari
Santosh Kumari
Santosh currently covers the Social Media and Politics beat for Attack of the Fanboy. She's been in the field of content writing for five years, yet she still can't resist checking for the next big story, even after her shift ends.