AI and Electronic Warfare Are Keys to Maintaining Air Superiority

I wanted to, I have two more quick questions here but the first is about AI initiatives. I think there's literally hundreds of AI programs across DOD right now but specifically how do you see this approving agility and mission success for the Air Force? Yeah, so I think when we talk AI and air superiority the, immediately jump to the flashy programs like Darpa Ace and Vista and Sky Borg. All the things are going to get us to CCA and the next-gen programs and those are very cool and making really impressive progress quickly but I think that we have to remember that AI is going to pervade everything. And with that in mind I think we need to focus on a couple areas additionally. So the first is open architectures. Air Force and DOD at large doesn't have the best track record of building open data systems particularly when we're trying to advance quickly. I think for AI we need to keep in mind balancing, building an open, scalable architecture that's going to evolve with the mission and the technology. The second is an area of trust. So, we're in the middle of the generative AI hype cycle. Companies spending billions of dollars on these models but they can't actually tell you why they work and generate all the results that they do. And, while that's fine if you have ChatGPT write a song about your dog in the style of Taylor Swift probably not perfect when we're doing intelligence and targeting decisions. So, need more work on there. And then the final area is Adversarial Encounter AI. We're not the only ones investing in these systems. We need to understand the vulnerabilities both to our own and the Adversaries. We've had experts starting to talk about things, techniques like data poisoning that could really defeat some of our AI-powered cyber defense systems. So, an area that needs more research and focus. 

I think we have to focus across all three pillars, attach, protect, and support. The standup of the new 350th Spectrum Warfare Wing is a great first step. I think that everybody agrees there’s more that needs to be done there and expect that there’s more that we’ll hear about that soon. That said, I think there are some really good efforts underway in terms of new programs. For example, the Angry Kitten platform, which is an AgilePod built by Georgia Tech Research Institute, flew in an unmanned aircraft earlier this year. AFRL's doing some amazing work in an adaptive and cognitive EW. And ACC Federal Labs is tackling both process and technology by showing how we can rapidly reprogram aircraft and weapons systems out at Beale, Nellis, and Edwards Air Force Base.