Reddit users demonstrated that searching for "site:claude.ai/share" surfaced numerous shared conversations covering topics ranging from programming and news discussions to more personal subjects. Reports also described exposed content that included health records, company documents, and personal information, although some claims circulating online, such as exposed cryptocurrency wallet information, could not be independently confirmed.
Anthropic moved quickly to address the issue. Google search results for the indexed Claude conversations were removed shortly after users highlighted the problem, although previously generated share links remained accessible to anyone who already had the URL.
Anthropic said shared conversations only appear in search engines if users publish those links somewhere that search engines can crawl, such as forums or social media. The company added that links shared privately are not discoverable through search.
"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," Anthropic spokeswoman Amie Rotherham said. "These shareable links are not guessable or discoverable unless people choose to share them themselves. 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 also emphasized that website owners control whether pages are indexed. "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," Google spokesperson Ned Adriance said. "We give site owners clear controls to decide whether pages can be crawled or indexed, and we always respect those directives."
The incident has renewed concerns about how AI companies handle shared content. Similar situations previously affected ChatGPT and Claude, where publicly shared conversations became searchable through Google after users generated share links.
For Claude users who have never shared a conversation, there is no indication that private chats were exposed. The issue applied only to conversations for which users created public share links.
Users can review existing shared conversations by navigating to Settings > Privacy > Shared Chats > Manage, where active share links can be disabled if they no longer want those conversations to remain publicly accessible.
The discovery also raises broader questions about how AI platforms communicate the implications of public sharing features. While Claude's sharing tool is intended to let users distribute conversations through a link, the indexing of those pages demonstrated that publicly accessible AI conversations can become searchable unless measures are in place to prevent search engines from cataloging them.
This analysis is based on reporting from ZDNET.
Image courtesy of Cybersecuritynews.
This article was generated with AI assistance and reviewed for accuracy and quality.
