A researcher has uncovered that almost 100,000 ChatGPT conversations, which users had marked for public sharing, were indexed by Google. This has created a dataset that exposes a wide range of personal and sensitive information, sparking concerns about privacy in the use of AI tools. The findings follow a previous report by Fast Company about thousands of shared ChatGPT chats appearing in Google search results, which has now been expanded with the new data. OpenAI confirmed the scale of the indexed chats but did not dispute the findings.