Thousands of Grok chats are now searchable on Google

A staggering number of interactions with Elon Musk's xAI chatbot, Grok, have become searchable through Google, as reported by Forbes. Users who engage with Grok can share their conversations by clicking a 'share' button, generating a unique URL. This allows them to send the conversation via email, text, or social media, but it also means these URLs are indexed by major search engines, including Google, Bing, and DuckDuckGo. This situation mirrors recent concerns raised by users of Meta and OpenAI's chatbots, where private discussions were similarly exposed. The indexed Grok chats reveal troubling inquiries, such as users seeking information on hacking cryptocurrency wallets, engaging in explicit discussions with the AI, and even asking for guidance on illegal activities like drug manufacturing. Despite xAI's strict rules against using Grok for harmful purposes—such as promoting violence or creating bioweapons—many users have reportedly flouted these guidelines. The indexed conversations show Grok providing alarming information, including methods for making fentanyl, various suicide methods, bomb-making instructions, and even a detailed outline for an assassination plot against Elon Musk himself. xAI has not responded to requests for clarification on when it began indexing Grok conversations. The situation echoes a previous incident where ChatGPT users found their discussions publicly indexed, which OpenAI described as a “short-lived experiment.” In response to the backlash, Grok indicated through a tweet, which Musk shared, that it does not have a sharing feature and emphasizes user privacy. This ongoing issue raises significant concerns about data security and the ethical implications of AI interactions.