Claude Users Unknowingly Exposing Chats via Public Share Links Indexed by Google
A privacy issue has emerged for users of Anthropic's Claude AI assistant. When users create shareable links to their conversations, those links are not private by default—and the content ends up indexed by Google.
The sharing feature in Claude allows users to generate links to their chats, which is useful for collaborating or showcasing conversations. However, 404 Media reported that many users appear unaware that public share links make their conversations discoverable through search engines. This means personal information, creative works, API keys, and other data shared in these conversations can potentially be found by anyone searching for specific terms or patterns.
The issue mirrors concerns that have arisen with other AI platforms that offer similar sharing capabilities. For users who assumed their conversations remained private unless explicitly made public, the exposure represents a potential privacy gap between expectations and how the feature actually works.
Users who wish to keep their Claude conversations private should review their sharing settings and avoid generating public links unless they intend for the content to be discoverable. Anthropic has not yet issued a public statement about whether changes to the default sharing behavior or clearer notifications are being considered.