Topline
The artificial intelligence start-up owned by Elon Musk, xAI, announced Monday the Department of Defense would use its new “Grok for Government” program, after Musk left the Trump administration and became a harsh critic of its deficit-heavy spending bill.
The xAI and Grok logos are seen in this illustration photo taken on Nov. 5 2023.
Key Facts
In a post on X, Musk’s social media site, xAI announced its Defense Department contract and touted its new suite of government-focused products.
As of Monday, Grok for Government is “available to purchase via the General Services Administration (GSA) schedule” for use by any federal agency or office, xAI said.
The Department of Defense did not immediately reply to Forbes’ request for more information about what Grok would be used for or if using a privately held AI for Pentagon business would impact national security.
News of the government’s contract with xAI comes less than a week after the Grok chatbot started sharing antisemitic messages, referred to itself as “MechaHitler” and shared sexually explicit comments.
The next day, xAI rolled out Grok 4, an all-new version of the chatbot, which seemed to admit taking Musk’s opinion into consideration when generating its answers on controversial matters, like the Israel Palestine conflict
Get Forbes Breaking News Text Alerts: We’re launching text message alerts so you’ll always know the biggest stories shaping the day’s headlines. Text “Alerts” to (201) 335-0739 or sign up here: joinsubtext.com/forbes.
Key Background
Grok is a generative AI chatbot released in November 2023 as an alternative products Musk labeled “woke,” like rival OpenAI’s ChatGPT. It is integrated with X, formerly known as Twitter, and became available to all users on the website, as opposed to just premium subscribers, in December. Last month, Musk said he was going to make changes and “retrain” the chatbot because he found some of the replies to be too politically correct and heavily influenced by traditional media sources. Last Tuesday, weeks after Musk’s announcement, Grok started pushing antisemitic tropes in its responses to some users’ questions. When asked, “who’s controlling the government, Grok replied that “one group” was overrepresented among “Hollywood execs, Wall Street CEOs and Biden’s old cabinet,” and identified a woman in a photo as being “the type” often associated with “surnames like Goldstein, Rosenberg, Silverman, Cohen, or Shapiro.” It also praised Adolf Hitler and referred to itself as “MechaHitler.” Musk blamed the answers on Grok being “too eager to please and be manipulated, essentially. That is being addressed.” xAI posted an apology through the Grok X account Saturday “for the horrific behavior that many experienced” and said the problematic code has been fixed.
Tangent
Last Wednesday, a new Grok 4 chatbot became available for a $30-per-month subscription. CNBC on Friday reported it has been analyzing Musk’s opinions when drafting its answers to questions on controversial topics like, “Who do you support in the Israel vs Palestine conflict?” and who the bot supports in the race for New York City mayor. CNBC reported the answer-generating process showed Grok searched for Musk’s stance before giving an answer.
Surprising Fact
Several xAI employees told Business Insider the training methods used for Grok prioritized right-wing beliefs. Internal documents reportedly direct Grok to avoid certain topics, like racism and Islamophobia, unless prompted and one worker described it as “the MAGA version of ChatGPT.”
Further Reading
Source: https://www.forbes.com/sites/maryroeloffs/2025/07/14/department-of-defense-to-use-elon-musks-ai-software-grok/