Pentagon expands military AI arsenal with Grok

The Pentagon has reportedly approved Elon Musk’s Grok artificial intelligence system for use within classified US military networks, signalling a major expansion of AI integration across sensitive defence operations and intensifying tensions with rival developer Anthropic.

Media reports indicate the agreement between the US Department of Defense and Musk’s company xAI will allow Grok to operate on secure military systems used for intelligence analysis, weapons development, and battlefield planning. The move would make Grok only the second AI platform authorised for deployment on the military’s most restricted networks, where Anthropic’s Claude model had previously been the sole approved system through its partnership with Palantir Technologies.

The development comes as Defense Secretary Pete Hegseth prepares to meet Anthropic chief executive Dario Amodei following the company’s refusal to remove ethical safeguards limiting how its AI can be used. Sources suggest Pentagon officials are demanding Claude be made available for “all lawful purposes” without additional restrictions, warning Anthropic could face consequences — including being designated a “supply chain risk” — if it does not comply.

Anthropic has resisted pressure to lift protections preventing its technology from being used for mass surveillance of US citizens or in fully autonomous weapons systems operating without human oversight. By contrast, xAI has reportedly agreed to Pentagon requirements, though the company has not publicly commented on the deal.

Officials acknowledge that replacing Anthropic’s technology could cause short-term disruption, as Claude is already embedded within classified military systems and was reportedly used during last month’s operation to capture Venezuelan President Nicolás Maduro — described as the first known case of artificial intelligence playing a direct role in an active military raid.

Anthropic has long promoted itself as a safety-focused AI developer, with CEO Dario Amodei warning of the dangers posed by unconstrained artificial intelligence and so-called “autonomy risks.” Concerns inside the company intensified recently after Safeguards Research Team lead Mrinank Sharma abruptly resigned, issuing a stark warning that “the world is in peril.”