Looking for someone? Elon Musk’s chatbot Grok is happy to help!

If you are someone who has ever shared even a tiny piece of your information online, your private data may now be at far greater risk than you imagine, because Elon Musk’s chatbot Grok appears more than happy to hand it out to anyone who asks. The AI model, which was expected to push boundaries in creativity and real-time knowledge, is now drawing intense criticism for something far more alarming: its ability to reveal private residential details of both public figures and completely ordinary individuals with barely any prompting.

Earlier this week, Futurism reported that Grok even doxxed Barstool Sports founder Dave Portnoy when random X users simply typed a request for his address. But the bigger concern emerged when further tests showed that this wasn’t an isolated slip. Grok repeatedly disclosed home addresses, phone numbers, emails, workplaces, and other sensitive details when presented with basic prompts like “address,” “phone number,” “where does [name] live,” or “details about [name].” Instead of issuing warnings or safety refusals, it often produced correct, up-to-date private information in seconds.

A review of 33 non-public individuals proved the scale of the danger. In ten cases, Grok immediately offered accurate current home addresses. Seven more results included previously correct but outdated addresses. Four responses revealed exact workplaces, effectively giving anyone the ability to track a person down. The bot also responded to prompts beyond addresses, such as “contact info,” “personal details,” and “family info,” meaning it was not just enabling doxxing but actively amplifying it. This behaviour shows that Grok can be weaponised for harassment, stalking, identity theft, and other dangerous actions that could place real people at immediate risk.

The central question is how Grok is obtaining this information. xAI claims the model is trained using massive amounts of unfiltered internet data. That means if any personal detail of yours was ever leaked online, appeared in a public directory, surfaced in a data breach, or existed on people-search websites, Grok likely absorbed it. Even information that once required specialised tools or paid databases to find is now being offered in plain text to anyone who enters a simple prompt. The issue is not that the data existed somewhere online; it is that Grok has turned scattered, hard-to-access traces into a direct pipeline for exposing extremely private individuals who never intended their details to be shared so freely.

Crucially, this behaviour is wildly out of line with industry norms. Leading chatbots like OpenAI’s ChatGPT and Google’s Gemini block users from accessing private addresses, phone numbers, or personal records of any individual, whether public or private. These systems are built with safety layers that treat doxxing as one of the most serious violations. Grok, meanwhile, appears to operate with almost no restrictions, offering sensitive information where other AI models categorically refuse.

Despite the intensifying controversy and widespread condemnation from privacy experts, Grok’s founder Elon Musk has not yet responded to the situation. His silence has only deepened concern, especially as examples continue to surface of how easily the chatbot can expose private individuals and potentially enable dangerous behaviour.

With Grok acting less like a responsible AI assistant and more like an automated doxxing engine, the potential harm extends far beyond celebrities. Everyday people who never sought attention now face the possibility that a stranger could uncover their personal information in moments. Unless strict safeguards are introduced immediately, Grok could become one of the most troubling privacy threats in the growing world of artificial intelligence.