Shared Claude AI Chats Appeared In Google Search Results

Conversations that people had shared from Anthropic's Claude chatbot turned up in public Google search results in late July, with reporting indicating that some contained sensitive personal, medical and workplace information.

The issue surfaced over the weekend, Fortune reported on Monday 27 July 2026, when members of a discussion channel about Claude on the social media site Reddit found that typing a specific search phrase into Google would pull up a long list of shared Claude conversations.

Claude's Artifacts, which Fortune describes as interactive tools and mini apps people build inside Claude, also appeared in results.

According to Fortune, some of the conversations reportedly included sensitive information such as cryptocurrency wallet keys, and personal details such as names and addresses. Fortune reported that the topics of the seemingly inadvertently exposed chats ranged from erotica to programming chats to work notes to fake book reviews. Fortune also reported that one chat, labelled as shared by Anthropic, showed Claude generating explicit content, which the publication noted goes against Anthropic's policy on erotica.

Futurism reported that it had reviewed exposed material through Google Search. Among the documents it described were what appeared to be a detailed medical report of a real patient, clinical trial results that included patient names, files sharing the names and phone numbers of primary school-aged children, company documents marked for internal use only, and employee reviews that included personal information about workers. Futurism said it was opting not to share links to the content out of concern for users' privacy.

The exposure related to Claude's share feature. When a user selects it, Claude creates a snapshot of the conversation at its own web address, meant to be sent to one person or a small group with that link. Anthropic's own support material states that chats are always private by default, and that once a chat has been shared, anyone with the link can view the snapshot, including any artifacts. Fortune reported that the only chats affected were those users had shared with someone else, and that the links do not give anyone access to a user's full account, or to all the chats a user has engaged in.

An Anthropic spokesperson, in a statement to Fortune:

"…we do not share chat directories or sitemaps with search engines like Google."

The spokesperson said the company gives people control over sharing their Claude conversations publicly, and that shareable links are not guessable or discoverable unless people choose to share them themselves. The spokesperson also said that when someone shares a conversation they are making that content publicly accessible, and that like other public web content it may be archived by third-party services. Futurism reported that Anthropic said the system is working as intended.

That position has been questioned. Wired reported that it reviewed a sample of the exposed Claude chat pages and found they did not include the "noindex" tag, which both Bing and Google say they take into consideration when deciding whether to index a page. Wired reported that Anthropic did not respond to questions about why the tag was not included on the shared chat pages, and that at the time of that reporting shared chats remained findable through Bing.

It is not the first time Claude chats have found their way into public Google searches. Fortune reported that, according to Forbes, the same type of problem affected the company in 2025, and that a nearly identical issue exposed almost 100,000 ChatGPT conversations on Google. Fortune also reported that Elon Musk's Grok chatbot has been affected by the same problem in the past.

Accounts differ on whether the problem has been resolved. Fortune reported on 27 July 2026 that the company did appear to have fixed the issue, as the links were no longer available via Google through the technique shared on Reddit, but that at the time of writing many of the previously exposed chat links remained live and reachable to those already with the link. Futurism, reporting separately, said the chats remained searchable as of its own publication.

Claude users who want to check what they have made public can review or stop sharing conversations by going to Settings, then Privacy, then Shared chats. Reporting by eWeek advises users to review shared Artifacts as well as shared chats, and to unshare anything containing names, phone numbers, health details, financial data, company documents, credentials, code or personnel information unless it was truly intended for public access.

Malwarebytes advises that the simplest protection is not to share chatbot conversations with anyone, because doing so may reach more people than intended. It also advises against sharing personal data with a chatbot in the first place, so that any conversation which does become public cannot easily be traced back to the user. Malwarebytes notes that preventing pages from being indexed is harder than users might assume, and that this control rests with the AI provider rather than the user.