The United States will allow the controversial xAI Grok bot to be used in its closed military systems

Summary of news about the U.S. Army’s collaboration with AI companies

The latest development is a new model for military AI: Elon Musk’s company agreed to provide its neural network Grok to the armed forces as part of secret systems. This became an official confirmation from the U.S. Department of Defense. Until today, the only AI model used by the armed forces for reconnaissance and combat operations was Claude from Anthropic. It had already been employed in a recent operation in Venezuela.

Anthropic’s requirements
The company demands assurance that its technologies will not be used for mass surveillance of U.S. citizens or for creating fully autonomous weapons. The Pentagon considers such conditions too restrictive and points out the presence of “gray areas” that make them unworkable.

Reaction from the Department of Defense
Defense Secretary Pete Hegseth almost severed ties with Anthropic, even declaring the company a threat to the supply chain. This means any contractor working with the department must end business relations with Anthropic.

Response from xAI
Unlike Anthropic, xAI agreed to use Grok “for all lawful purposes.” However it is still unknown whether this model can fully replace Claude and how long such a transition would take. The Department of Defense acknowledges that replacement is not an easy task.

Negotiations with other players
The Pentagon has accelerated talks with OpenAI and Google, preparing for a possible break in relations with Anthropic. According to sources:
• OpenAI is still far from finalizing a deal;
• Google is closer to an agreement allowing the use of Gemini in secret projects.

Position of the Department of Defense
An official spokesperson rejected New York Times’ claims, stating that negotiations are underway with both companies and the department expects contracts to be signed provided the “all lawful purposes” criterion is met.

Key takeaways
1. xAI is now officially part of the Army’s AI arsenal—Grok can be used in all secret systems if legal requirements are satisfied.

2. Anthropic remains under pressure due to its restrictions; the Department of Defense has nearly severed ties with it.

3. The Pentagon actively seeks alternatives, speeding up negotiations with OpenAI and Google to avoid dependence on a single company.

Thus, U.S. armed forces are expanding their AI resources while continuing to face regulatory and security challenges in the technology domain.