When Elon Musk's Grok AI had a Nazi meltdown the other month and started calling itself "MechaHitler," the General Services Administration — the agency in charge of government technology — quietly dropped its plans to use it by removing it from a list of approved vendors.

Now, it looks like the White House has swooped in to change that. Wired reports that the executive branch has ordered GSA leaders to deploy the notoriously foulmouthed chatbot across the federal workforce as quickly as possible. 

"Team: Grok/xAI needs to go back on the schedule ASAP per the WH," reads an email obtained by Wired sent by Josh Gruenbaum, commissioner of the Federal Acquisition Service. "Can someone get with Carahsoft on this immediately and please confirm?" he added, referring to a contractor that resells tech from third-party firms to the government.

It appears that the call has been answered. As of Friday, both Grok 3 and its recently released upgrade Grok 4 — which Musk boasted will discover "new technologies" and even "new physics" just days after the chatbot self-identified as its creator to explain why he'd visited the home of notorious sex criminal Jeffrey Epstein — now appear on GSA Advantage, an online marketplace where federal agencies shop for products and services. Sources also told Wired that the Carahsoft contract had been updated earlier this week to include xAI, Musk's AI company that offers the chatbot.  

It's a pretty fortuitous reversal of fortunes. As Wired previously reported, xAI lost out on a massive government contract that the GSA had made with its competitors like OpenAI and Google after Grok suddenly went off the rails in one of the most spectacular meltdowns in the history of the AI industry. In a psychotic posting spree, the bot espoused Nazism, shared rape fantasies, hinted at a second Holocaust, and memorably styled itself as "MechaHitler."

The deal was on the verge of going through before that, and Grok was even added to the GSA's Multiple Award Schedule, its long-term contracting program, in late June. Once its edgelord posting spree made headlines a little over a week later, though, it was removed from the contract offering — but not straight away. GSA employees were baffled at how, at first, their higher-ups were still strangely enthusiastic about deploying the AI, fresh off of its Nazi crash out.

"The week after Grok went MechaHitler, [GSA leadership] was like 'Where are we on Grok?'" a GSA employee told Wired in previous reporting. "We were like, 'Do you not read a newspaper?'" 

It wouldn't be until early August that Grok would be kicked to the curb.

That the decision has now been overturned is hardly surprising. It was more was more surprising that an agency under an openly "anti-woke" Trump administration would turn up its nose up an AI model being racist in the first place, for the brief period that it did press "pause." Since Trump took office, numerous agencies have embraced using AI to help conduct its purge of DEI and any other traces of "woke" content and policies in government. The administration blamed AI for deleting a page dedicated to Black baseball legend Jackie Robison,for instance.

So, yes. We suppose it does follow that it's apparently of paramount importance that the government grunts over whom the sword of DOGE hangs from a single strand of horsehair get their hands on the "I love Adolf Hitler" AI without delay. 

More on AI: Man Suffers ChatGPT Psychosis, Murders His Own Mother


Share This Article