AI JOE

How to train a model for war.

- WILL KNIGHT

AT THE START of this year, the newly restyled US Department of War proclaimed that the American military “will become an ‘AI-first’ warfighting force,” “unleashing” AI agents on everything “from campaign planning to kill chain execution.” Less than two months later, the Trump administration unleashed its AI-fueled attacks on Iran—and some of the agents balked. Department officials went head-to-head with Anthropic executives over the terms of a roughly $200 million contract. The company wanted to limit the use of its models in autonomous weapons. Defense Secretary Pete Hegseth was having none of it and declared Anthropic a supply-chain risk—prompting the company to file a pair of federal lawsuits. While Anthropic might have misgivings about providing the US military with unfettered access to its AI models, some startups are building advanced AI specifically for military applications. Smack Technologies, which announced a $32 million funding round in early March, is developing models that it says will soon surpass Claude’s capabilities when it comes to planning and executing military operations. And, unlike Anthropic, the startup appears less concerned with banning specific types of military use.

“When you serve in the military, you take an oath you’re going to serve honorably, lawfully, in accordance with the rules of war,” says CEO Andy Markoff. “To me, the people who deploy the technology and make sure it is used ethically need to be in a uniform.”