Claude shared chats indexed by Google, risking sensitive info leaks

Security firm Malwarebytes found that shared chat links from Anthropic's AI assistant Claude can be indexed by Google, potentially exposing sensitive info from users' chats. The issue was first reported by Reddit users, who found many shared Claude chats through specific searches.

The problem comes from Claude's 'Share chats' feature, which creates public links for conversations. Malwarebytes says some public links get crawled and indexed by search engines. Similar issues have happened with Grok and Meta AI.

To stay safe, Malwarebytes advises users not to share AI chats publicly and to avoid entering personal details, passwords, or confidential info in conversations.