Anthropic Claude Conversations Exposed in Google Search Results Incident
Key point
A security incident occurred in which conversations from Anthropic's Claude users who used the 'share' feature were exposed in Google search results.
Details
A security issue occurred in which conversations from Anthropic's chatbot Claude were exposed in Google search results. This issue appeared in chats where users used the 'share' feature, which is used to pass along conversation records to others.
It was reported that the search results included everything from sensitive personal information such as cryptocurrency wallet keys, names, and addresses to programming code, work notes, and even adult content that violates Anthropic's policies. In addition, content from Artifacts, Claude's interactive tool, was also exposed in the search results.
This issue was first revealed through a Claude-related community on Reddit, and it was confirmed that entering certain search terms would surface a large number of shared Claude conversations.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.