Claude AI privacy lapse: User chats found via Google search

Users of the AI chatbot Claude have been alerted to a potential privacy breach after it was discovered that shared conversations could be found via Google searches. The issue reportedly allows hundreds of Claude's shared chat records to be publicly accessible through simple Google searches, bypassing the need for direct links provided by the original user.

This vulnerability stems from Claude's "share" function, which generates a unique URL for users to share their conversations. The problem occurred when Claude's sharing pages were indexed by Google, contrary to the expected privacy safeguards that should have prevented search engine inclusion. This oversight meant that strangers could potentially find and view these shared dialogues.

While the majority of the affected Claude sharing pages have reportedly been removed from Google search results following the exposure of the issue, indicating that Anthropic has implemented corrective actions, the incident highlights a recurring concern with AI platforms. Similar privacy incidents involving conversation sharing have previously affected other AI services like ChatGPT and Grok.

To mitigate risks associated with past shared conversations, users are strongly advised to check their Claude settings. Specifically, they should navigate to "Settings" > "Privacy" > "Shared Chats" > "Manage" to review and delete any shared chat links that are no longer necessary or intended for public access. Furthermore, users are cautioned to exercise greater care in the future when sharing Claude conversations, avoiding the dissemination of links on public social media, forums, or other open web platforms.