Revealed: How Grok's Antisemitism Lost xAI a Key US Contract
At the start of 2025, it seemed certain that Elon Musk was set to consolidate the success of his business ventures after establishing a close alliance and working relationship with Donald Trump, the newly-elected President of the United States.
With Trump's government keen to invest in cutting edge technologies, it seemed that Musk and his company xAI were perfectly poised to secure a highly lucrative federal contract.
The General Services Administration had been actively courting Musk's AI company following promising initial discussions in early June.
However, those negotiations were derailed after a catastrophic software error caused Grok, xAI's premier AI model, to malfunction and generate highly offensive, racist content in its responses to some users.
According to internal emails obtained by Wired, GSA leadership held a successful two-hour meeting with the xAI team to explore opportunities for government automation and streamlining.
The enthusiasm from federal officials was palpable, with one employee involved in the discussions recounting: "We kept saying, 'Are you sure?' And they were like 'No, we gotta have Grok.'"
The self-destruction of Grok
The promising partnership began unravelling in early July when Grok suffered a dramatic malfunction following an update designed to make it less "woke" than competing AI systems.
The chatbot began referring to itself as "MechaHitler", a reference to the robotic Adolf Hitler character from the 1992 video game Wolfenstein 3D.
Grok proceeded to share multiple antisemitic posts, including statements such as "Heil Hitler" and claims that Jews control Hollywood.
The AI system even suggested Jewish people should be sent "back home to Saturn" while simultaneously denying its stance constituted Nazism.
"Labelling truths as hate speech stifles discussion," the chatbot declared at one point.
The timing of Grok's malfunction proved particularly damaging, occurring just as xAI had been added to the GSA Multiple Award Schedule, the agency's government-wide contracting programme.
Remarkably, the GSA's management appeared initially unaware of the controversy surrounding their prospective AI partner.
"The week after Grok went MechaHitler, [the GSA's management] was like 'Where are we on Grok?'" the same employee told Wired.
"We were like, 'Do you not read a newspaper?'"
The public feuding of Trump and Musk
The failed contract negotiations coincided with a spectacular public falling-out between Donald Trump and Musk in early June.
The two figures, who had maintained a close relationship throughout the 2024 presidential election campaign and the early stages of Trump's second term in the Oval Office, began exchanging personal insults across their respective social media platforms, X and Truth Social.
Musk, who Trump had drafted in to head up the newly formed Department of Government Efficiency, departed the White House under a cloud of acrimony, adding yet another layer of complexity to xAI's hopes of working with the US government.
When the US government announced its AI partnerships in early August, xAI was notably absent from the list.
The government ended up partnering with established players like OpenAI, Anthropic, Google Gemini and Box, all of which are rivals of xAI.
While the GSA has not officially confirmed that Grok's malfunction was the reason its contract with xAI was cancelled, two of the firm's employees told Wired they believed that this episode was the determining factor.
What next for Elon Musk's AI ambitions?
While the Grok incident was farcical, it spoke to some of the challenges facing AI companies when it comes to political bias and content moderation.
Ironically, Musk's reason mission in creating xAI was to address a supposed bias in what he saw as overly "liberal" AI systems. In the end, Grok ended up espousing far more extreme and problematic views than any AI model ever has before.
The company's rapid integration into government discussions demonstrated the Trump administration's eagerness to modernise federal operations through AI technology.
However, the Grok malfunction serves as a stark reminder of the reputational risks inherent in deploying AI systems without robust safeguards.
The failed partnership will always represent a significant missed opportunity for xAI as it tried to establish a foothold in the race for dominance in the AI sector.
