Christopher Kirchhoff, tech adviser to multiple presidents, on Air Force One

Christopher Kirchhoff is co-author of “Unit X: How the Pentagon and Silicon Valley are Transforming the Future of War” takes readers inside the AI race for national security and private industry’s critical role. As the co-founder of the Defense Innovation Unit with his co-author Raj Shah, he created the Pentagon’s first permanent advanced tech outpost based in Silicon Valley and is assisting the current Department of Defense on technology efforts. In a new interview with host Tammy Haddad on the Washington AI Network Podcast, he gives a compelling case for the urgency to invest in next-generation technology, the threats from China, Russia and North Korea and lays out how the Trump Administration can drive technological advancement forward to meet a new set of global threats. 

“The greater danger is not that AI itself will pose a danger to us, but that we won’t experiment with AI in national security applications,” Kirchhoff said. He described how hesitancy to embrace experimentation could leave the U.S. vulnerable to adversaries who are accelerating their technological capabilities.

Kirchhoff pointed to the war in Ukraine as a vivid example of how modern warfare is evolving. “We gave Ukraine 31 M1A1 Abrams battle tanks, the most advanced battle tank in the world. And in short order, in the last few months, the Russians have figured out how to engineer kamikaze drones. And today, more than half of those M1A1 tanks have been taken off the battlefield by inexpensive Russian drone strikes,” he explained. This, Kirchhoff argued, highlights the growing importance of adaptability and innovation in modern defense strategies. “That tells a new incoming administration that if you are running a military that’s built with a lot of tanks, you now are behind the power curve and you need to find new ways of experimenting with new weapons platforms to be able to compete.”

When asked about the Pentagon’s efforts to address these challenges and the urgency of now, Kirchoff replied saying, “we don’t have, unfortunately, 10 years to figure out how to master and deploy this new set of technology. Our adversaries are already hard at work, as we can see in Ukraine, in the Middle East, in the Red Sea, in China, and so we really have got to go fast and bring the best parts of these technology ecosystems together,” Kirchhoff said. He emphasized the need for urgency, explaining that initiatives like Replicator are essential for staying ahead of global competitors.

Kirchhoff also discussed the role of private-sector leaders like Elon Musk in advancing defense innovation. “Elon would like it if we called him the X factor in the incumbent administration, and he really is,” Kirchhoff said. He praised Musk’s contributions to autonomous systems and satellite networks, particularly their impact on conflicts like Ukraine. However, Kirchhoff cautioned against over-reliance on private actors. “The Department of Defense is not a private company. You can’t just fire three-fourths of the people that are there and start over. You have to work within the existing system,” he noted.

When asked about the changing defense landscape, Kirchhoff described the current moment as an inflection point. “We’re in a remarkable moment that its many parts collectively constitute an emergency because those collective parts tell us that our existing force structure, our existing military technologies are not able to defend powerfully and muscularly against these new threats, that it’s time to change the technology and the operating concepts that we’re using to fight,” he said.

When asked about the cultural shifts required to embrace new technologies, Kirchhoff called for a bold rethinking of how defense institutions operate. “The government model is that you’re not willing to pay technologists market rate, you’re not going to get any of the good technologists working in the government,” argued Kirchoff. “Now, if you’re a taxpayer, I would want a really great set of technologists working in the government, helping the government figure out which technology to buy. And that problem is still there. We’re still at a moment in the middle of this artificial intelligence revolution where not at a single department or agency of the government can actually afford to hire an AI engineer at market rate.”

Closing the conversation, Kirchhoff delivered a clear call to action: “We’re a long way, I think, from having to worry about AI posing that level of threat. So the greater danger is not that AI itself will pose a danger to us, but that we won’t experiment with AI in national security applications. The AI safety community in their enthusiasm to make things perfectly safe I think have actually prevented an experimentation with the national security community that would’ve happened earlier,” he said. 

I think the biggest hurdle that we face is having the right agenda for technology modernization, but having the wrong kind of politics to see it through, in the sense that we’re about to embark upon quite the national squabble on a lot of political fronts. And the challenge is in that kind of moment, that kind of political moment, whether you can actually get Congress to move a reform agenda through, or whether there’s going to be no bandwidth or serious reform at a moment of real political wrangling.

For Kirchhoff, collaboration across the public and private sectors is critical to meeting the challenges of the 21st century. “I think whoever begins to helm the Department of Defense in this new administration should work very closely and cooperatively with all the frontier labs to put those models to work and see where they can perform a really useful function and to make sure they’re deployed in contexts that are safe and not deployed in contexts that those models can’t support, that they’re not ready for,” he concluded.