Claude shared chats and Artifacts found publicly visible on Google — privacy alert for AI users

Hundreds of Claude conversations and interactive Artifacts — mini applications and documents built inside Anthropic’s AI assistant — were publicly indexed by Google over the weekend, exposing sensitive personal and business data. Reddit users discovered the leak by using the search operator “site:claude.ai/share,” which pulled up a long list of shareable links to chats. Some of these contained health records, internal company documents, and the names and phone numbers of school-aged children, according to reports from Futurism and Fortune.

The incident has reignited concerns about data privacy in AI collaboration platforms, particularly for Indian businesses and professionals who rely on Claude for document drafting, code generation and team workflows.

What actually happened

The exposure stems from Claude’s share-chat feature, which allows users to generate a public link for any conversation or Artifact. The interface warns: “Anyone with the link can view.” However, unlike Google Docs’ similar sharing model — where indexed documents do not automatically appear in search results — Claude’s public links were found to be easily crawlable by search engines once those links were posted anywhere on the public web, such as forums, social media or embedded on third-party sites.

By Monday afternoon, a test search by TechCrunch using the same operator no longer returned results, suggesting Anthropic had already taken steps to block crawling — likely by adding a noindex directive or updating the site’s robots.txt. It is not yet clear whether all exposed links were removed before they could be cached or archived.

Anthropic and Google respond

Anthropic has largely placed responsibility on users. Spokeswoman Amie Rotherham told TechCrunch: “We give people control over sharing their Claude conversations publicly, and in keeping with our privacy principles, we do not share chat directories or sitemaps with search engines like Google. 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.”

Google echoed a similar stance. Spokesperson Ned Adriance said: “Neither Google nor any other search engine controls what pages are made public on the web, and these pages were indexed across many search engines. We give site owners clear controls to decide whether pages can be crawled or indexed, and we always respect those directives.”

The company noted that the same issue affected multiple search engines, not just Google, and that site owners — in this case Anthropic — have the tools to prevent indexing.

Past patterns and scale

This is not the first time Claude’s shared conversations have leaked. In 2024, Forbes reported a similar episode where hundreds of chats were indexed by Google; the company estimated just under 600 conversations were exposed before being removed. Whether this weekend’s leak is larger or smaller has not been independently confirmed.

Parallel issues have affected other AI platforms. In 2024, 404 Media reported that a researcher scraped around 100,000 ChatGPT conversations that had been set to public. The pattern suggests a recurring design challenge: making collaboration links shareable yet not automatically discoverable by search engines.

What Indian users should check now

Anyone who has used Claude’s share-chat feature should review their shared conversations immediately. Go to Settings → Privacy → Shared Chats to see which links are active. For each one, decide whether to delete the link or change its visibility. If you shared a chat link on any public platform — a blog, a forum post, a LinkedIn comment — assume it may have been indexed.

For enterprises in India handling sensitive data, this incident is a reminder to restrict use of shareable links for internal conversations and to audit third-party AI tools for their data-handling defaults.

Key data

Item	Details
Event	Claude shared chats and Artifacts found on Google via site: operator
Date	First reported July 27, 2026 (Saturday to Monday)
Data exposed	Health records, internal company files, children’s personal info, explicit content
Company response	Anthropic says users made links public; no index sitemap was shared

Source: TechCrunch — PSA: your Claude shared chats and Artifacts may have ended up on Google (https://techcrunch.com/2026/07/27/psa-your-claude-shared-chats-and-artifacts-may-have-ended-up-on-google/)