Hundreds of conversations with the popular AI chatbot Claude were found publicly available online. Users shared links to their chats, and search engines like Google saved these links. As a result, anyone could find and read the private talks. The conversations covered more than 200 talks across 25 pages of search results. Many users did not expect their private talks to become public, and they felt their privacyprivacy/ˈprɪvəsi/L2隐私；私密the state of being free from public attention or being watched by others was not protected.
The chats included personal information like names, work projects, and private questions. For example, one user asked Claude to turn them into a fox, while another shared details of a secret work project. Some people even used the chatbot to help with their resumes and for proprietaryproprietary/prəˈpraɪəteri/L2专有的；专利的owned or controlled by a particular company or person, often not shared with others work research. A spokeswoman for Anthropic said, 'Links to conversations are not guessable or discoverablediscoverable/dɪˈskʌvərəbəl/L2可被发现的；可被找到的able to be found or found out unless people choose to share them themselves.'
The company quickly blocked the links from search results, and similar issues happened before with other AI chatbots. A Google spokesman said they give site owners clear controls for web pages. The event teaches a valuablevaluable/ˈvæljuəbəl/L2有价值的；宝贵的very useful or important lesson about online privacy. If you share a link, it might become public in ways you did not expect. You should always think carefully before sharing private information online, and this is a good reminderreminder/rɪˈmaɪndər/L2提醒；提示something that makes you remember or think about something to be careful on the internet.
