An American AI Action Plan w/Charles Clancy and Joshua Levine

While Silicon Valley builds advanced AI models and Beijing integrates them into state power, Washington faces an uncomfortable reality: America's innovation machine might not be enough to win the AI race on its own. The problem isn't our technology—it's our government's ability to deploy it.

The White House recently released “America’s AI Action Plan,” which aims to change this dynamic, calling for everything from "Manhattan Project-style" coordination to federal AI sandboxes. But with the Trump Administration now moving to implement these initiatives, the question becomes: can American democracy move fast enough to compete with authoritarian efficiency? And should it?

Charles Clancy, Chief Technology Officer of MITRE, knows the challenges well. His organization serves as a bridge between government needs and technical solutions, and he’s seen firsthand how regulatory fragmentation, procurement bottlenecks, and institutional silos turn America's AI advantages into operational disadvantages. His team also finds that Chinese open-weight models outperform American ones on key benchmarks—a potential warning sign as the U.S. and China compete to proliferate their technology across the globe.

Clancy argues the solution is not for the U.S. to become China, but rather to take a uniquely American approach—establish federal frontier labs, moonshot challenges, and market incentives that harness private innovation for public missions. He and FAI’s Josh Levine join Evan to explore whether democratic institutions can compete with authoritarian efficiency without sacrificing democratic values. View Mitre’s proposals for the White House’s plan here, and more of Charle’s research here.