PSA: Your Claude shared chats and Artifacts may have ended up on Google

TechCrunch reported on July 27 that Anthropic's Claude AI assistant has a significant privacy issue with its share chat feature — conversations shared via public links may have been crawled and indexed by Google Search.

The issue originates from Claude's share chat feature, which allows users to create links that enable anyone with the assigned URL to view a conversation or project. However, TechCrunch found that these shared pages were apparently indexed by Google, making private conversations appear in search results.

This means conversations that users believed were only accessible through a specific link could actually be found by anyone searching for relevant keywords. For conversations involving business secrets, personal information, or sensitive topics, this poses a serious data exposure risk.

Anthropic has not yet issued an official response. TechCrunch has reached out to Anthropic for comment, and it remains unclear how many conversations are affected or the full scope of Google's indexing.

This is not the first time AI chat tools have faced privacy scrutiny over sharing features. Similar concerns have been raised about ChatGPT's sharing mechanisms. While sharing links are designed for collaboration, users often underestimate the risk of those pages being indexed by search engines.

For Claude users, the immediate recommendation is to audit any previously created shared links for sensitive information and exercise caution with the share feature. Enterprise users in particular should reassess their data security policies around Claude's sharing functionality.

The incident has reignited discussion about privacy-by-design in AI products. Security experts argue that AI companies should default to blocking search engine indexing on shared pages and provide clear privacy warnings at the point of sharing.