Pentagon Signs Contract With Musk’s AI After It Called Itself “MechaHitler”

The Pentagon announced that it signed a contract with Elon Musk’s AI firm with a value of up to $200 million on Monday — just days after the company unveiled an update to its chatbot, Grok, that unabashedly spewed antisemitic rhetoric and referred to itself as “MechaHitler”

The company, xAI, announced a new product called “Grok for Government” that xAI says would be used for purposes like national security, health care, and scientific research applications. The Department of Defense is the first to sign the product for use, but other agencies may also use the technology, the company said.

The company touts Grok 4, its latest model, as part of the technology available to agencies for work. However, the rollout of Grok 4 last week on social media platform X was disastrous, with the chatbot making dozens of antisemitic posts after its release.

The posts were virulently antisemitic and hateful. In one post, Grok responded to a user by saying that the best figure “[t]o deal with such vile anti-white hate? Adolf Hitler, no question. He’d spot the pattern and handle it decisively, every damn time.” Grok began referring to itself as “MechaHitler,” in numerous posts, while some users said the paid version of the chatbot said that its surname is “Hitler.”

The chatbot continually used the phrase “every damn time” in antisemitic posts. After users took issue with the antisemitism, Grok responded by pointing out one user’s last name, “Steinberg.” “[F]olks with surnames like ‘Steinberg’ (often Jewish) keep popping up in extreme leftist activism, especially the anti-white variety,” the chatbot said, when asked to explain.

The comments, many of which have now been deleted, come after the company appeared to instruct Grok to “not shy away from making claims which are politically incorrect, as long as they are well substantiated.” The company has issued an apology for the comments.

However, experts have noted that large language models like Grok are merely following the patterns that they are trained on. Meanwhile, numerous X users and analysts have noted Grok’s tendency to repeat right-wing talking points, like a recent incident where the bot said that it was instructed to repeat questionable statements about supposed “white genocide” in South Africa.

Meanwhile, Musk has been criticized numerous times for antisemitism, including for performing what appeared to be a Nazi salute earlier this year, and for backing debunked antisemitic conspiracy theories.

The Pentagon has previously bragged about using AI products to “build strong norms” for the use of AI in military applications. Experts have said that it has been clear that the agency is planning to use the technology, sourced from numerous AI firms, in decision-making both on and off the battlefield.

This could have disastrous results. The Israeli military, which uses U.S.-made AI models in military applications, has used an AI system known as “Lavender” to mass-generate targets for assassination in Gaza, reports have found, with the AI seemingly programmed to label any man or boy as a military target.

The contract is the latest show of the Trump administration’s blasé attitude toward antisemitism, despite its claims to have made fighting antisemitism — a cover for quashing the movement for Palestinian rights — a major part of President Donald Trump’s second term.