Hundreds of conversations with Anthropic’s AI chatbot Claude were found publicly accessible through Google and other search engines, BBC News reported on July 27. The chats — some containing personal information, work details, and what appeared to be proprietary research — were discoverable by anyone using site-specific search terms.
🔍 THE BOTTOM LINE
The exposure was not a hack. Users who clicked Claude’s “share” button created public links to their conversations, and search engines indexed those links. Anthropic has since blocked the chat logs from search results, but the incident echoes nearly identical problems at OpenAI’s ChatGPT and xAI’s Grok — raising the question of whether AI chatbot share features are systematically producing privacy exposures across the industry.
What Happened
According to the BBC report, links to Claude conversations showed up in Google search results when users searched using site-specific terms. The shared chats included:
- A user asking Claude to draft an unpublished blog post about cloud security involving details of a corporate project
- Users seeking resume help, including their names, contact information, and work history
- Someone conducting what appeared to be proprietary healthcare research, including transcripts of private conversations
- A user asking how to “become a Nine-tailed fox,” with Claude responding by generating an AI image claiming they had been given “fully functional fox powers”
Reddit users initially discovered the publicly available chats, which covered more than 200 conversations across at least 25 pages of search results.
Not a Breach — But Not Fine Either
An Anthropic spokeswoman told the BBC that Claude users maintained control over whether to share conversations, and that links were “not guessable or discoverable unless people choose to share them themselves.” She said that when someone shares a conversation, “they are making that content publicly accessible, and like other public web content, it may be archived by third-party services.”
The share feature within Claude tells users that “anyone with the link” may view the contents, but does not explicitly state that the link may end up in Google and search results. The search availability was removed over the weekend, but many conversations were saved and shared widely online before the fix.
An Industry Pattern, Not an Anthropic Problem
This is not the first time an AI chatbot’s shared conversations have surfaced in search results. OpenAI experienced an almost identical issue with ChatGPT chat logs being made publicly accessible last year, and ultimately changed how easily such logs could be found. Grok, the AI chatbot within Elon Musk’s X platform, also saw hundreds of thousands of chat logs made publicly available through online search.
The pattern points to a structural issue: AI chatbot share features create public URLs by design, and search engines index public URLs by design. The gap between those two design decisions and user expectations of privacy is where the exposure lives.
A Google spokesman told the BBC that the company does not control “what pages are made public on the web,” and that site owners have tools to block crawling and indexing. In other words, the responsibility falls on the platform — not the search engine.
What Anthropic Did About It
The search indexing of the chat logs was stopped over the weekend, likely through Anthropic using available tools to block the chat log links from search results. Google’s process for a website owner to block a link is straightforward but must be initiated by the website owner — in this case, Anthropic.
Other search engines including Bing, Brave, and DuckDuckGo, through which the Claude chat logs also appeared, were approached for comment by the BBC.
NZ Angle
For New Zealand businesses using Claude or similar AI tools, the lesson is straightforward: treat every shared AI conversation as a public document. If you paste corporate research, client details, or internal strategy into a chatbot and then share the link — even with one colleague — that conversation can end up indexed by search engines. NZ’s Privacy Act requires reasonable steps to protect personal information, and sharing an AI chat link containing personal data could create compliance exposure if that data becomes publicly discoverable.
❓ FAQ
Was Claude hacked? No. Users chose to share their conversations using Claude’s share feature, which created public links. Search engines then indexed those links.
Did Anthropic fix it? Yes. The search indexing was removed over the weekend, likely through Anthropic blocking the chat log links from search results using standard webmaster tools.
Have other AI companies had this problem? Yes. OpenAI’s ChatGPT and xAI’s Grok both experienced nearly identical issues with shared chat logs appearing in search results last year.
Should I stop using Claude’s share feature? Not necessarily, but treat any shared chat as a public document. Avoid sharing conversations that contain personal information, corporate secrets, or anything you would not want indexed by Google.
🔍 THE BOTTOM LINE
The incident is a reminder that “share” buttons in AI chatbots are not private messaging tools. They create public web pages, and public web pages get indexed. Anthropic, OpenAI, and xAI have all now learned this the same way — through user conversations turning up in search results. The fix is always reactive, and the next platform to discover it will likely do so the same way.
