Anthropic Defends Claude Chat Privacy After Shared Links Indexed on Google

A simple search query over the weekend exposed conversations between users and Anthropic’s Claude AI assistant on Google, laying bare medical reports, company files, children’s phone numbers, and cryptocurrency wallet keys. The incident sparked widespread attention across the tech press after a Reddit user flagged the search operator site:claude.ai/share on a Saturday, returning a long index of user-generated links. 404 Media first reported it on Monday, and the rest of the tech press followed within hours.

Anthropic’s position is that nothing failed. The company asserted that the system, it says, is working as intended.

Anthropic Defends Privacy Design as Shared Links Surface on Google

Claude conversations are private by default. Only shared ones were affected. When a user shares a conversation, sharing creates a snapshot of the chat at its own public web address, meant for a colleague or a small group. The problem is what “public” turned out to mean. A simple “noindex” tag tells search engines to leave a page out of results. It does not appear Anthropic used one.

Anthropic spokeswoman Amie Rotherham addressed the exposure by emphasizing user control and privacy principles. Amie Rotherham told TechCrunch that they give people control over sharing their Claude conversations publicly, and in keeping with their privacy principles, they do not share chat directories or sitemaps with search engines like Google.

Amie Rotherham also stated that those shareable links are not guessable or discoverable unless people choose to share them themselves, and that when someone shares a conversation, they are making that content publicly accessible.

Anthropic spokeswoman Amie Rotherham

The company added that links only reach search results once a user posts them somewhere a crawler can see. In effect, it placed the responsibility with the people who pressed share.

Comparing Platform Controls and Exposure Risks

That defence runs into an obvious counter-example. Google Docs offers a near-identical share feature. Those documents do not end up searchable on Google. The difference is the missing instruction to search engines.

Google’s spokesperson Ned Adriance told TechCrunch that site owners hold the crawl and index controls, and that Google respects them. Consequently, the choice not to use them rested with Anthropic.

The exposed material was not trivial. Futurism reviewed the results through Google. It found a detailed medical report on a named patient, clinical-trial results listing patient names, and documents holding the names and phone numbers of primary-school-aged children. Other cases were worse for being ordinary. Reddit users cited a person building a crypto wallet who exposed its keys. Another was a lawyer asking Claude whether they had to self-report a breach of professional conduct. Claude Artifacts were caught in the same net. These are the interactive apps and documents people build in the tool. Internal dashboards and project plans with client data turned up alongside the chats.

Broader Industry Vulnerabilities in Generative AI

AI company Anthropic announced Tuesday that it is moving its popular Claude Cowork agent to the cloud, making it accessible to users from multiple devices, and able to perform tasks even when the devices are not online. Cowork is a feature that was built primarily for nontechnical users. Launched in January, it automates multistep workflows for personal organization and scheduling by pulling from users’ files, folders and applications without the need for constant prompting or direction. Up to now, Cowork had only been available for users operating on specific devices, which needed to remain open and active for work to be completed, and users could not operate Cowork via multiple devices using one account.

The update announced Tuesday will change that. The changes announced will allow scheduled tasks, like drafting an email, to run without a device online, but Claude will still ask a user for a final review and approval before shipping, the company said in its blog. Anthropic will give beta access to users with Max subscriptions, the company’s most premium plan for consumers, and will expand access to more plans in the coming weeks. To mark the launch, Anthropic is extending Cowork’s doubled usage limits through Aug. 5.

Users concerned about past activity can manage their exposure by navigating to Settings, selecting Privacy, opening Shared Chats, and revoking any active links.