According to tech media outlet 404 Media, a large amount of user conversation content and creative works generated while using Anthropic’s Claude AI assistant were publicly indexed by Google Search. This means anyone searching for relevant keywords could potentially view these private conversations.
The report indicated that exposed content includes users’ private conversations with Claude, as well as articles and project content created using Claude. The leakage pathway is not yet fully clear, but may be related to default configuration settings of certain public sharing features on the Claude platform.
Anthropic has not yet issued a formal statement on this incident. Cybersecurity experts note this incident reflects significant vulnerabilities in user privacy protection on AI platforms, particularly regarding data sharing and search engine crawler blocking. As AI tools become increasingly integrated into daily life, user data protection issues urgently require industry attention.
Editor: GoodInfo Global News Team