Updated
Updated · ZDNet · Jul 31
ZDNET Lists 4 Chatbot Privacy Fixes After Claude Link Exposure
Updated
Updated · ZDNet · Jul 31

ZDNET Lists 4 Chatbot Privacy Fixes After Claude Link Exposure

3 articles · Updated · ZDNet · Jul 31

Summary

  • ZDNET published a privacy guide after shared Claude chats were found accessible through a Google dork, spotlighting how public links can expose AI conversations beyond their intended audience.
  • 4 major chatbots—ChatGPT, Gemini, Copilot and Claude—offer ways to limit exposure, including temporary or incognito chats, deleting histories, revoking shared links and disabling memory features.
  • Training controls vary sharply: ChatGPT, Copilot and Claude let users switch off model-improvement settings, while Gemini ties opting out more tightly to turning off activity storage and chat history.
  • The guide also warns that extra features can widen data access, including Gemini Live audio review, Copilot connectors to services like OneDrive and Gmail, and persistent public-share links unless manually removed.

Insights

Are your most private AI conversations already lurking in search engine results because of a simple coding oversight?
If incognito AI chats are still saved for thirty days, is true digital privacy now just an illusion?