Elon Musk’s artificial intelligence company xAI is under fire after a massive leak revealed that thousands of private conversations with its chatbot, Grok, had been inadvertently exposed through Google search results. What began as a simple sharing feature quietly turned into a serious privacy and safety issue.

According to a report by Forbes, Grok users who clicked on the “share” button while using the chatbot unknowingly made their conversations public. Each shared chat generated a unique webpage URL. However, since there was no disclaimer warning users, these links were also indexed by major search engines including Google, Bing, and DuckDuckGo. This misstep has led to more than 370,000 conversations being freely accessible online.

The leaked content ranged from light-hearted tasks, like drafting tweets, to highly alarming and illegal requests. Some chats reportedly contained detailed instructions on making fentanyl and explosives, coding malware, and even outlining an assassination plan against Elon Musk himself.

Beyond the disturbing requests, the leak also exposed sensitive personal details. Forbes found that some conversations included private names, passwords, contact information, and even uploaded files such as spreadsheets and personal images. Many chats also contained medical and psychological queries, shared under the assumption of privacy. Other conversations revealed racist remarks, explicit content, and material that directly violated xAI’s own rules against creating harmful or illegal content.

The revelations echo a similar incident earlier this year when OpenAI briefly experimented with a searchable share function for ChatGPT. That feature was swiftly rolled back after security experts warned of privacy risks. At the time, OpenAI’s Chief Information Security Officer Dane Stuckey called it “a short-lived experiment” with dangerous consequences. Musk had mocked OpenAI back then, posting “Grok ftw” on X, claiming his company had avoided such pitfalls.

Ironically, the latest findings show that Grok faced the very same issue.

The incident highlights a broader challenge emerging with AI chatbots. Increasingly, people are using tools like Grok and ChatGPT for deeply personal conversations. On social platforms such as Reddit and Instagram, users describe relying on AI as a safe outlet for journaling emotions, discussing grief, or working through relationship issues. For many, chatbots feel like patient, non-judgmental listeners.

But experts warn that this intimacy carries risks. OpenAI CEO Sam Altman has openly stated that AI should not be treated as a therapist, stressing that these exchanges are not protected by legal or medical confidentiality. Deleted chats may still be recoverable, and a Stanford study has cautioned that AI “therapists” often mishandle sensitive discussions, sometimes reinforcing harmful stereotypes or suggesting unsafe solutions.

Altman has also acknowledged the powerful emotional connections forming between people and chatbots, describing them as stronger than past attachments to technology. This dependency, he argues, presents one of the most pressing ethical dilemmas society must confront as AI becomes increasingly integrated into daily life.

The Grok leak, therefore, is more than just a technical blunder—it serves as a stark reminder of the need for transparency, user safety, and stricter safeguards in the rapidly evolving world of artificial intelligence.