Chatbots are designed in such a way that you feel your conversation is personal. They’re much more agreeable and friendlier, and mimic the experience of chatting with a friend online—assuming the friend can instantly generate a wall of text in a matter of seconds. In reality, however, they are far from personal, especially if your definition of “personal” includes “for your eyes only.”
Cloud Chats was visible on Google this weekend
Some cloud users can attest to that fact after the weekend. As reported by Cyber Security NewsAnthropic’s cloud share links appear in Google search results, indicating that search engines like Google have indexed users’ chats. on weekends, A viral Reddit post Just adding “site:claude.ai/share” to your search queries, highlighted how you can pull up real cloud chats directly into Google.
If a cloud user talked about a keyword you searched for and Google indexed the conversation, you can pull it up and read the transcript. According to Cyber Security News, it includes lawyers discussing legal strategies, engineers talking through technical problems and users being open with their personal problems. Depending on the conversation, it can range from mildly embarrassing to downright embarrassing.
Now, some caveats: Anthropic quickly nipped the issue in the bud after the Reddit thread started making waves. If you try to search for anything with “site:claude.ai/share” now, you’ll get a failed search result instead, so all your Cloud chats should be safe going forward. Perhaps most importantly, it does not appear that these indexed chats are affected Everyone which uses the cloud; Instead, these were “shared links,” which the cloud generates when you share a conversation with someone else. The flaw seems to stem from this feature – if you’ve shared a cloud chat, it could end up in this searchable database.
That said, not all leaked chats may fit that criteria. A similar situation happened in Septemberand, As reported by FuturismAt least one affected user refused to share their chat. True to their word, it’s possible that some of the open cloud chats are from users who never used the Share Chat function.
AI chats are not private
If you’re sure your cloud chats weren’t shared, or believe a leak has been plugged, that doesn’t mean your AI chatbot conversations should be treated with a presumption of privacy. Even if you have never shared a cloud conversation, and thus may not be affected by this issue, your chats will not be included in your account. While you might think that what you say to your AI stays between you and the chatbot, everything you say to your bot is is shared with Anthropic and used to train future models—unless you explicitly opt out.
What do you think so far?
It’s not just anthropic: most, if not all, AI companies operate this way, and each has its own set of policies. Some individuals can review your chat directly, for example—you might not think of asking your AI bot for help with a personal problem.
Even if you opt out of data logging, the company in question may hold your chats for a certain period of time. Anthropic, for example, Your chat data will be retained for 30 days After you delete the conversation. So not only is data held on Anthropic’s servers in the first place, it stays there after you can’t access it. Also, Anthropic isn’t the only company dealing with indexing exposure: while this is the second time this has happened to cloud users, ChatGPT users faced a similar problem last August.
If you choose to use chatbots, Never share personal or private information with them-Especially Highly sensitive data like health information. AI is undoubtedly here to stay, and many find it useful, but if you care about your privacy, these tools simply don’t have your best interests in mind.
