Over the weekend, it was discovered that many Claude chats and Artifacts—interactive mini apps and documents created within Anthropic’s AI assistant Claude—were publicly searchable on Google. Users on Reddit found that using specific Google search queries like “site:claude.ai/share” brought up numerous shared conversations, some of which included sensitive information such as health records, private company documents, and personal details of children. These exposures stemmed from Claude’s “share chat” feature, which allows users to generate links for sharing conversations or projects, but these links inadvertently became accessible on the open web.

Anthropic responded by emphasizing that shared links are only indexed by search engines if users post them in publicly accessible places like forums or social media. The company clarified that sharing a link via private means should keep content out of search results. Anthropic also reiterated that their platform does not provide direct access to chat directories or sitemaps for search engines, so content is only discoverable if users intentionally make it public. Despite this, the incident raised concerns due to the nature of the exposed material, which included not only personal data but also content that violated Claude’s usage policies, such as sexually explicit material.

The issue first came to light through a Reddit post on Saturday and was further reported by 404 Media on Monday morning. TechCrunch verified the exposure but later confirmed that Google appears to have removed the shared Claude links from its search results by Monday afternoon. Previous similar incidents, including one last year where hundreds of Claude conversations were indexed by search engines, show a pattern of recurring privacy risks tied to AI chatbot platforms’ sharing features. Users are advised to check their shared content settings within Claude and manage their privacy preferences accordingly.

Google clarified that search engines do not control what content is made public online but respect site owners’ directives on indexing. Anthropic warned users that sharing any Claude conversation publicly carries the risk of the content being archived or indexed externally by third parties. Given the nature of the revealed data, this incident underscores the importance of cautious use of AI sharing features and the need for improved safeguards to prevent unintended public exposure of sensitive AI-generated content. Wend’s Markets followers should be mindful when sharing AI conversations and review the privacy settings on platforms like Claude regularly.