The U.S. government is apparently readying a major deal with Elon Musk's xAI, meaning the controversial Grok will soon be privy to some of the country's biggest secrets. Grok has ballooned from its initial November 2023 release, billed as a rival to ChatGPT, as Musk called out his AI competitors. The world's richest man claimed that ChatGPT was being trained to be 'politically correct' and had originally planned to call his model TruthGPT as he complained: "The danger of training AI to be woke – in other words, lie – is deadly."
This move away from being politically correct has seen Grok come under fire amid issues where it referred to itself as MechaHitler, became obsessed with white genocide, and was also accused of asking minors for 'nudes'.
Everything boiled over when Grok issued an apology after officials confirmed it had manipulated images of children as young as 11.
Advert
Governments in countries including the United Kingdom have gone as far as they could ban X, while India and Malaysia have followed through with banning Grok altogether.
As some countries pull back from Elon Musk's AI creation, a landmark deal is set to bring Grok closer to the bosom of the USA than ever before.
The New York Times reports how Anthropic CEO Dario Amodei has been summoned to Washington over concerns that the artificial intelligence giant won't lift restrictions that would make Claude available for "all lawful purposes." Antropic has reportedly clapped back, saying that it's willing to loosen restrictions under the promise that guardrails are put in place to stop its own AI from being used for the likes of mass surveillance or making autonomous weapons.
Despite the Defense Department signing a $200 million pilot contract with Anthropic in 2025, it's been looking further afield to xAI and Google's Gemini. Sources told the outlet that these deals are hoping to put pressure on Amodei and Anthropic to loosen their grip.
This might come as no real surprise, as speaking at Texas' SpaceX headquarters in January 2026, Defense Secretary Pete Hegseth said Grok's integration into military systems would be live by the end of the month: "Very soon we will have the world’s leading AI models on every unclassified and classified network throughout our department."
This is part of a supposed 'AI acceleration strategy' where the Department of Defense will "unleash experimentation, eliminate bureaucratic barriers, focus on investments, and demonstrate the execution approach needed to ensure we lead in military AI and that it grows more dominant into the future." Axios says xAI has agreed to an "all lawful use" standard for Grok on Pentagon systems, although it's not clear whether Grok will fully be able to replace Claude or how long it would take. We're reminded that a partnership with Palantir meant Anthropic's tech was used in the January 2026 raid on Venezuela that captured President Nicolás Maduro.
Officials admit that offloading Claude and replacing it with Grok would be a 'difficult' process, although a meeting between Hegseth and Amodei presents the latter with an ultimatum that threatens a "supply chain risk" and other penalties.
Whereas Grok, Gemini, and ChatGPT are all available in the U.S. military's unclassified systems, there are accelerated talks to move them all into the classified space as the pressure against Anthropic mounts.
UNILADTech reached out to xAI for comment.