Full story
Claude links indexed
A privacy incident involving Anthropic’s Claude AI emerged after shared chats and Artifacts were indexed in Google Search, with the exposure tied to Claude’s share chat feature that generates public URLs for conversations.
“Récemment, des conversations tenues avec l'intelligence artificielle Claude se sont retrouvées indexées dans les résultats de recherche de Google”
Simon Cousineau said the issue is not a data leak or a bug but “une mauvaise utilisation de la fonction de partage,” explaining that when a user clicks “Partager,” the conversation becomes public by default and can be indexed by Google.
A Reddit user’s discovery described in Decrypt found that a Google search for “site:claude.ai/share” returned hundreds of other people’s conversations, and the cause was “one missing line of code” because Anthropic’s share feature lacked a noindex tag.
The same Decrypt report said Anthropic’s robots.txt blocked crawling but “blocking a crawl doesn’t prevent indexing if the URL appears elsewhere on the web,” meaning the pages could still be cataloged once the links were posted publicly.
What users saw
The Decrypt report said the thread briefly became a public record of what people felt comfortable saying to an AI, including a lawyer asking whether they needed to self-report a professional ethics violation and a person pasting a crypto wallet seed phrase.
VentureBeat reported that shared Claude Artifacts—including interactive applications, dashboards, documents and other AI-generated work products—were also appearing in Google Search results, and it said it could not access any shared conversations while verifying that some Artifacts were searchable and accessible without authentication.
VentureBeat also quoted an Anthropic spokesperson saying, “We give people control over sharing their Claude conversations publicly,” and added that “These shareable links are not guessable or discoverable unless people choose to share them themselves.”
ZDNET framed the issue as a way to “see if your conversations were exposed,” stating that “Private chats were not leaked,” and it advised users to be cautious about creating share links and to check Settings > Privacy > Shared Chats.
Fixes and remaining risk
Cryptonews.net reported that Anthropic and search engines Google, Bing, and DuckDuckGo de-indexed the major leak affecting Claude customers, but it said search engines “forgot to take down a similar, inadvertent data dump” from Perplexity Computer files.
“A Reddit thread revealed this weekend that a basic Google search gave access to hundreds of conversations shared on Claude, Anthropic's conversational assistant, a figure some forum users were estimating in the thousands”
Cryptonews.net said Google began pulling Claude results by July 26 after Anthropic updated its configuration and added noindex tags, while also noting that “more specific queries with various keywords might return additional content.”
The Decoder said users can review and manage shared conversations under Settings -> Privacy -> Shared Chats, and it warned that the “Share with link” feature lacked a noindex tag so shared pages could be indexed by search engines.
Futurism reported that after hitting the “share” button in the top right corner, Claude notes that sharing will create a “public” link where “anyone with the link can view,” and it said the warning does not alert users that shared content could wind up being indexed by a search engine.
