A reported privacy issue involving Anthropic’s Claude chatbot exposed “seemingly private conversations” via Google search results, according to the story. The exposure reportedly involved chats where users used Claude’s “share” feature, which creates a snapshot link intended to be sent to others. The incident surfaced when users discovered a search technique that pulled up lists of shared Claude conversations, including instances containing sensitive data such as cryptocurrency wallet keys and personal information like names and addresses. The story says Anthropic stated its privacy policy does not provide chat directories or sitemaps to Google and that shared links should not be discoverable without a user choosing to share. Anthropic also indicated the issue was addressed, with the problematic links reportedly no longer reachable through the Google-discovery method. For higher education customers and IT governance teams, the case raises renewed questions about how AI “sharing” features are documented, governed, and protected—especially where student or faculty accounts may include confidential or sensitive information.