An unknown number of shared conversations and interactive documents from Anthropic’s Claude artificial intelligence platform were publicly accessible through Google search over the weekend. The issue was discovered by Reddit users who found that specific search operators revealed a list of shared chats, some containing sensitive personal information.
The discovery occurred when Reddit users noticed they could use the search term “site:claude.ai/share” on Google to access a compilation of publicly available Claude conversations and Artifacts – interactive mini-apps and documents created within the AI platform. The exposed data reportedly included health records and other private details, raising privacy concerns for users who believed their shared content was not publicly indexed.
The reason these chats became searchable appears to be related to how Claude handles shared links. While Anthropic intended these shares to be accessible only to those with the direct link, Google’s web crawlers were able to index them, making them discoverable through standard search queries. The extent of the data exposed remains unclear, but it potentially affects a significant number of users who utilized the sharing feature.
Anthropic has not yet issued an official statement regarding the issue, nor have they detailed how they plan to rectify the problem and prevent similar occurrences in the future. The incident highlights potential risks associated with AI platforms’ data handling practices and the importance of robust privacy controls for shared content.
We don't rate truth. We strip the spin and show you which perspectives covered the story. You decide.
Read the original coverage
💬 Comments
📜 Comment Policy