Recent findings have unveiled that conversations and creations from Anthropic's Claude AI are searchable on Google, leading to the exposure of sensitive personal information, including private cryptocurrency wallet keys and addresses. The data surfaced includes various applications, such as an AI therapy app and a dashboard for medical billing analysis, highlighting the potential for serious data breaches stemming from user sharing settings. Anthropic has stated that the responsibility lies with users who intentionally make their conversations public, emphasizing that they do not index chat directories or sitemaps for search engines.
For businesses utilizing AI platforms like Claude, this incident underscores the critical importance of understanding and managing data sharing settings and privacy controls. Organizations must educate their employees about the risks associated with sharing sensitive information through AI tools, especially in environments where data could inadvertently become public. This incident serves as a cautionary tale about the implications of user-generated content and the need for robust cybersecurity practices to safeguard sensitive data against unintentional exposure in an era where AI applications are becoming increasingly integrated into business workflows.
---
*Originally reported by [Schneier on Security](https://www.schneier.com/blog/archives/2026/08/some-claude-chats-are-searchable-on-google.html)*