Hundreds of Claude AI chats exposed via Google indexing
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic's Claude shared chat links were indexed by Google, Bing, Brave, and DuckDuckGo, making hundreds of user conversations accessible via site-specific searches before Anthropic blocked indexing over the weekend.
- Reddit users initially surfaced the publicly available chats, which spanned more than 200 conversations across at least 25 pages of search results, with some posted just weeks ago.
- Claude's share feature warns users that 'anyone with the link' can view the conversation but does not explicitly tell users the link could appear in Google search results, an Anthropic spokeswoman acknowledged.
- Exposed conversations included an unpublished corporate cloud-security blog post, resumes containing names and contact details, and transcripts of what appeared to be proprietary healthcare research.
- OpenAI's ChatGPT and Elon Musk's Grok on X both experienced nearly identical indexing problems last year, and OpenAI ultimately changed how easily such logs were accessible — a fix Anthropic has not yet replicated.
- Google told the BBC it does not control 'what pages are made public on the web,' saying the decision comes from websites and that site owners have clear tools to block crawling or indexing.
Why it matters: Anthropic's share feature tells users 'anyone with the link' can view a chat — but doesn't warn that search engines like Google will index those links, leaving users who thought they were sharing private URLs exposed. OpenAI changed ChatGPT's behavior after a similar 2024 incident; Anthropic has only blocked search indexing, not redesigned the share flow.
