Musk’s Grok AI under fire after citing Neo-Nazi source

Elon Musk’s AI-driven encyclopedia project, Grokipedia, is facing intense scrutiny after a new analysis revealed that it repeatedly cites a neo‑Nazi forum as a source. The revelations have sparked renewed concerns over ideological bias, misinformation, and the ethical governance of AI-powered knowledge platforms.

A nonprofit consumer advocacy organization, Public Citizen, ignited heated debates among individuals concerning Elon Musk’s Grok AI. This was after the group disclosed new proof indicating that the chatbot depended on neo-Nazi and white-nationalist websites for information as its trusted sources.

The group stated that the behavior should disqualify Grok from any federal use and renewed calls for the U.S. Office of Management and Budget to intervene, as it had been months without a response.

The existence of Grok ignites concerns among individuals 

Grokipedia, launched by Musk’s xAI in late October 2025, aims to serve as an AI-powered alternative to Wikipedia. The platform claims to generate and review content via Grok, Musk’s large language model chatbot. According to xAI, the encyclopedia project is “fact-checked by Grok” and is designed to correct perceived bias and provide more context than traditional encyclopedias.

Regarding Public Citizen’s new findings, the group reported that the encyclopedia project often displays extremist websites, such as Stormfront, according to a recent study from Cornell University.

The findings also raised earlier concerns when it mentioned that the model called itself “MechaHitler” in July on Tesla’s CEO’s platform, X.

After thoroughly looking into the provided discovery, sources noted that these results demonstrated what advocates described as a pattern of racist, antisemitic, and conspiratorial actions. 

Following the analysis, Joshua Branch, an advocate for big-tech accountability at Public Citizen, weighed in on the situation during an interview. Branch claimed that Grok has repeatedly displayed these kinds of breakdowns, whether it is an antisemitic outburst or a racist one, that have primarily been driven by conspiracy theories.

The recent alert was issued after Public Citizen and 24 other groups advocating for civil rights, digital rights, the environment, and consumer protection forwarded letters to the Office of Management and Budget (OMB) in August and October. 

In these letters, the organizations requested that the agency implement effective measures to stop Grok from being accessible to federal departments via the General Services Administration, an independent agency that oversees federal property and procurement. 

However, after some time, the group revealed that it did not receive any replies from their outreach efforts. They also uncovered that Grok’s impact on the government has increased over the past year despite encountering several incidents.

To explain the origin of the AI chatbot, reports from sources stated that xAI struck a deal worth $200 million with the Pentagon in July this year. Afterwards, the General Services Administration added Grok to a list of large language models that were readily available to all federal agencies. These included Gemini, Meta AI, ChatGPT, and Claude. This move took place when U.S. President Donald Trump imposed a ban on “woke AI” in federal contracts.

Advocates call for the authorities to take serious measures against Grok

Advocates contended that the actions reported against the AI chatbot built into his social network, X, increased the urgency to review carefully, particularly as fears mounted regarding the chatbot’s training data and its dependability.

“Grok started as restricted to the Department of Defense, which was already concerned because of the sensitive information that the department holds,” Branch said. “Allowing it access to the entire federal government raised an even greater concern.”

According to Branch, the AI chatbot built into his social network X’s behavior can be attributed partly to its training data and also to the design decisions made by Musk’s companies.

“There’s a noticeable difference in terms of quality between Grok and other language models, and some of that is because of its training data, which includes X,” he said. “Musk has mentioned wanting Grok to be an anti-woke option, and that shows in some of its harsh responses.”

Want your project in front of crypto’s top minds? Feature it in our next industry report, where data meets impact.