Main summary
The emerging artificial intelligence company owned by Elon Musk, XAI, announced Monday that the United States Department of Defense will use its new program “Grok for Government”, after Musk was distanced from the Trump administration and harshly criticize its draft spending law with high deficit.
Key data
In an X publication, the social network of Musk, XAI announced its contract with the Department of Defense and promoted its new range of products focused on the government sector.
As of Monday, Grok For Government is “available for purchase through the General Services Administration Schedule (GSA)”, allowing its use by any federal agency or office, XAI said.
That same day, the Department of Defense also reported that it granted contracts worth up to 200 million dollars to XAI, Google, Anthropic and OpenAi, in order to accelerate the adoption of advanced artificial intelligence capabilities to face key national security challenges.
The Department of Defense did not immediately respond to the request for Forbes information on what Grok will be used or if the use of a private property AI for pentagon tasks could represent a risk for national security.
The news of the contract between the government and XAI arrives less than a week after the Chatbot Grok began to disseminate anti -Semitic messages, “Mechahitler” was defined and published comments of explicit sexual content.
The next day, XAI launched Grok 4, a completely new version of the Chatbot, which seemed to admit that Musk’s opinion takes into account by generating answers on controversial issues, such as the conflict between Israel and Palestine.
Context
Grok is a generative artificial intelligence chatbot launched in November 2023 as an alternative to products that Musk described as “Wake”, as OpenAi chatgpt. It is integrated into X (before Twitter) and, since December, it is available for all site users, not only for premium subscribers.
Last month, Musk said he would make changes and “wake up” the chatbot because he considered that some answers were too politically correct and influenced by traditional means. Last Tuesday, weeks after the announcement, Grok began offering answers with anti -Semitic stereotypes.
When asked “Who controls the government?”, Grok replied that “a group” was overrepresented between “Hollywood executives, CEO of Wall Street and the old Biden cabinet”, and identified a woman in a photo as the type “that usually has surnames like Goldstein, Rosenberg, Silverman, Cohen or Shapiro.”
He also praised Adolf Hitler and defined “Mechahitler.” Musk attributed the responses to the fact that Grok was “too complacent and manipulable.” XAI published an apology in the Grok account in X on Saturday for “the horrible behavior they experienced” and said the problematic code was corrected.
Additional data
Last Wednesday, Grok 4 was launched, available via a subscription of $ 30 per month. According to CNBC, Chatbot has been analyzing Musk’s opinions by generating answers on controversial issues, such as “Who do you support in the conflict between Israel and Palestine?” or about his position in the election by the Mayor’s Office of New York. CNBC reported that the response process showed that Grok first looked for Musk’s posture before answering.
Surprising data
Several XAI employees told Business Insider that the training methods used for Grok prioritized right -wing beliefs. Internal documents indicated that Grok should avoid certain issues such as racism or Islamophobia, unless he was specifically asked, and one of the workers described him as “the magic version of Chatgpt.”
This article is a translation of the United States Forbes article
Do you like to inform yourself for Google News? Follow our showcase to have the best stories
