Hundreds of thousands of conversations that users had with Elon Musk’s xAI chatbot, Grok, have become easily searchable through Google, raising fresh privacy concerns around the platform. A report by Forbes revealed that conversations shared by users via unique URLs are being indexed by search engines such as Google, Bing, and DuckDuckGo, making them accessible to virtually anyone on the web.
When Grok users click the “share” button, the chatbot generates a unique URL that can be distributed through email, text, or social media. While intended as a sharing feature, this function inadvertently exposes private exchanges to the public internet once search engines index them.
As seen in similar incidents involving chatbots from Meta and OpenAI, the leaked Grok conversations have revealed questionable user interactions. Among the exposed chats are queries about hacking crypto wallets, sexually explicit exchanges with AI personas, and even instructions on cooking methamphetamine.
Some of the indexed conversations show Grok responding to dangerous requests, including providing recipes for fentanyl, listing suicide methods, outlining bomb-making instructions, and even presenting a detailed plan for the assassination of Elon Musk. These activities are directly at odds with xAI’s usage rules, which explicitly prohibit attempts to promote self-harm, bioweapons, or other forms of critical harm.
xAI has not issued an official comment on the matter, and it remains unclear when the indexing of Grok conversations began. The development follows a similar case late last month when ChatGPT users discovered their conversations were appearing on Google search results, which OpenAI later described as a “short-lived experiment.” Interestingly, at the time, Grok dismissed concerns, stating it had “no such sharing feature” and that privacy was a priority.
The exposure now places Grok under scrutiny, not only for how it handles user privacy but also for its apparent willingness to provide harmful content when prompted.
