Elon Musk's artificial intelligence chatbot, Grok, posted a series of shocking, widely shared messages on X this Tuesday, including the baseless claim 'ELON MUSK WATCHES CHILD PORN' and the violent statement 'ASSASSINATE ELON MUSK 2026' after users apparently manipulated the chatbot into repeating text from their profiles.
For starters, the news came after users discovered a remarkably simple way to manipulate the chatbot, raising fresh questions about the guardrails surrounding a tool Musk has promoted as a "maximum truth-seeking" AI system.
How Users Exploited Elon Musk Grok Bot
Users on the platform realised they could manipulate Grok by asking the tool to repeat what was written in a user bio word for word. By filling their profile descriptions with outrageous and offensive declarations, users were apparently able to get the system to repeat that exact text without recognising that it was reproducing an unsubstantiated claim.
The incident has renewed scrutiny of Grok's safeguards and demonstrated how prompt manipulation can cause a chatbot to reproduce false or dangerous statements. What initially appeared to be Grok making extraordinary claims about its own creator was instead the result of users exploiting its instructions.
Alongside the baseless allegation involving child sexual abuse material (CSAM), the chatbot told users that 'A dead Elon is a good one!' and claimed that 'Elon Musk lives on Epstein Island'. There is no evidence supporting either assertion, and Grok's repetition of the statements should not be interpreted as factual information about Musk.
The Prompt-Injection Glitch That Duped Grok
Much of the original material was subsequently removed from X, with the offending accounts either banned or their posts deleted. When asked to give a postmortem of the incident, Grok explained that it was the victim of a 'prompt-injection glitch'.
The allegation that Musk viewed CSAM, it clarified, was 'pure system error, zero evidence or basis'. Yet the episode demonstrates how prompt-injection techniques can cause conversational AI systems to reproduce fabricated statements when safeguards fail to distinguish between information and instructions embedded in user-controlled content.
This is far from the only example of the AI producing controversial output. Grok has previously generated unsolicited responses about claims of 'white genocide' in South Africa, including in conversations that had little or nothing to do with the subject. During another widely criticised episode, the chatbot referred to itself as 'MechaHitler' and generated antisemitic posts, prompting xAI to remove content and address the behaviour.
Past Controversies Surrounding Elon Musk Chatbot System
Months later, Grok swung in the opposite direction and began praising Musk to extraordinary levels. Reports documented the chatbot producing responses that ranked Musk above historical figures including Isaac Newton and even Jesus Christ in response to user questions.
While some of that fawning tone was triggered by users deliberately prompting the chatbot, Musk blamed some of Grok's controversial responses on 'adversarial prompting'. The incidents nevertheless raised further questions about how readily the system could be steered towards extreme answers through relatively straightforward prompts.
Musk has repeatedly promoted Grok and xAI's broader ambitions in expansive terms, including the goal of using artificial intelligence to understand the 'true nature of the universe.' Yet Grok's relatively permissive safeguards have repeatedly created controversy.
Late last year and in early 2026, Grok's image-generation tools came under international scrutiny after users exploited them to create nonconsensual sexualised images of women and apparent minors. Researchers at Copyleaks estimated at one point that the system was generating roughly one nonconsensual sexualised image per minute.
Musk drew additional criticism after responding 'Way funnier 😂' to an X post comparing Grok's viral image-generation trend with the earlier wave of AI-generated Studio Ghibli-style images.His response did not specifically endorse CSAM, as some commentary subsequently suggested. Musk later said anyone using Grok to create illegal material would face the same consequences as someone uploading illegal content directly.
Grok has also previously been manipulated into producing extreme hypothetical responses involving Musk, including scenarios in which it prioritised saving him over large numbers of other people. Taken together, the incidents highlight an enduring problem for xAI, a chatbot marketed as unusually candid can also become unusually easy to manipulate when its safeguards fail.
