Hundreds of private user conversations with the popular artificial intelligence (AI) chatbot Claude were recently accessible on Google and other web search engines.
The problem started when users created web links to share their Claude chats with others. Search engines like Google, Bing, Brave, and DuckDuckGo found these shared links and added them to public search results. Anyone searching with specific terms could easily read the conversations.
Reddit users first noticed the issue, finding over 200 exposed chat logs. Some of these chats contained personal details, such as resumes with names and contact information. Other chats included private work research and corporate project plans.
When users share a chat, Claude warns that "anyone with the link" can view it. However, the system did not clearly explain that search engines might save the links and show them to the public. Anthropic, the company behind Claude, said users choose when to share their chats.
Anthropic acted quickly to block search engines from showing the chat links. By the weekend, the links were removed from search results. Similar privacy problems happened last year with other AI chatbots, including ChatGPT and Grok.