Claude's chat sharing feature sparks privacy concerns as Google indexes conversations
Claude's chat sharing feature sparks privacy concerns as Google indexes conversations
A privacy scare hit Anthropic's Claude AI after users discovered that shared chat links were being indexed by Google, exposing sensitive information like health records and corporate documents. The incident, first spotted on Reddit, highlights the risks of public sharing features in AI tools. Anthropic says the issue stems from users posting links publicly, but critics argue the design encourages oversharing.
What happened?
Last Saturday, a Reddit user noticed that by using Google search operators like site:claude.ai/share, they could pull up a trove of publicly shared Claude conversations. Some of these chats contained sensitive data: health records, internal company documents, children's names, and phone numbers. The discovery quickly spread, and by Monday morning, tech outlet 404Media had reported on it.
The problem lies with Claude's "Share Chat" feature, which lets users create public links to their conversations. While the interface warns that "anyone with the link can view it," the design seems intended for sharing with a small group—friends, colleagues, or a team—not for broadcasting across the entire internet. But once a link is posted on a forum, social media, or any site that search engines crawl, it becomes fair game for indexing.
Anthropic's response
Anthropic was quick to clarify that the company doesn't proactively submit chat links to search engines. Spokesperson Amy Rotherham explained that the links only appear in search results if users themselves publish them on publicly accessible platforms. "After users choose to publicly share, the content can be indexed by third-party services like other public online content," she said.
But that explanation hasn't fully satisfied critics. The exposed content includes not just casual conversations but also code, work notes, medical reports, clinical trial documents, and employee reviews. In some cases, real patient information and children's contact details were visible. Even more awkwardly, Fortune reported that one shared chat contained explicit content generated by Claude—despite Anthropic's usage policy explicitly banning such material.
A familiar problem
This isn't the first time AI chat privacy has made headlines. Last year, Forbes reported that hundreds of Claude chat records had been indexed by search engines, with Google estimating about 600 conversations in its index. Around the same time, 404Media covered researchers scraping roughly 100,000 publicly shared ChatGPT chats.
The pattern is clear: as AI chat tools become more integrated into daily work—handling everything from medical advice to corporate strategy—the mechanisms for sharing and controlling access are struggling to keep up.
What's being done?
By Monday afternoon, TechCrunch found that the previously public content was no longer retrievable via the same search method, suggesting that the issue had been partially addressed. Google noted that it cannot control which web pages websites choose to make public; website owners can use mechanisms like robots.txt to prevent crawling and indexing.
For users worried about their own Claude chats, Anthropic offers a way to check: go to Settings → Privacy → Share Chat to see which conversations have been shared publicly.
Key Points
- Privacy breach: Claude's shared chat links were indexed by Google, exposing sensitive data.
- Root cause: The "Share Chat" feature creates public links; when users post them on accessible sites, search engines can index them.
- Anthropic's stance: The company does not proactively submit links; indexing occurs only when users share links publicly.
- Historical context: Similar incidents have occurred with Claude and ChatGPT in the past.
- Resolution: The issue appears to have been partially resolved, but the underlying design remains a concern.
