🌊 Breaking Down AI Military Integration

How is AI integration in the military changing warfare?

On January 2, 2026, airstrikes hit Venezuela, knocking out electricity in the capital, Caracas. Under the cover of darkness, Delta Force operators descended on the presidential compound in Caracas, secured then-president Nicolás Maduro and his wife, and extracted them to a US ship stationed in the Caribbean. The whole operation lasted only two and a half hours.

It was also the first time the US used a generative AI model – Anthropic’s Claude – in a military operation.

The operation was the result of nearly a decade’s worth of AI integration in the US military. While the exact details of Claude’s role in the Venezuela operation aren’t fully clear, the military has been using the technology widely in the Middle East. Simultaneously, it’s been battling AI companies over how, exactly, it can use their products: To surveil Americans? To order lethal strikes?

In today’s deep-dive, we examine how the military has integrated AI, how it’s using platforms like Claude in Iran, and what safeguards – if any – exist around it.

In April 2017, the Department of Defense (DoD) established the Algorithmic Warfare Cross Functional Team – known unofficially as Project Maven or simply Maven – to speed up the integration of AI algorithms across the US military.

Maven initially focused on using AI to identify potential military targets by processing drone and satellite footage, thus reducing the burden on human analysts. By automating detection, labelling, and classification of certain targets, analysts could focus on making higher-level decisions rather than spending hours staring at screens.

The DoD first contracted Google to help develop the algorithms behind Maven, and in late 2017, special forces operations in East Africa first used the program. But internal backlash at Google over the ethical implications of AI integration in the military forced the company to abandon the contract.

Enter, Palantir.