White House Office of Science and Technology Policy Director Michael Kratsios | Global Conference

Great. So OSTP is one of the four policy offices in the White House. There's an economic policy office, a domestic policy office, the national security, and science and tech. What's very interesting about science and technology policy is that it's spread out across so many agencies. You have the National Institutes of Health doing health work. You have National Science Foundation doing science work. You have the Department of Defense building and investing in high-end technologies for the warfighters, and so many other agencies. What we do at the White House is coordinate policy across all of those agencies. So when the President says we must win in artificial intelligence, how do we drive that strategy to make sure that all of our agencies are doing what's necessary for each of them in their domains to make sure we achieve the President's mission?

Well, first and foremost, the President wants us to win on these key critical and emerging technologies. We, as a White House, have spoken about four in particular: artificial intelligence, quantum information science, nuclear energy, and biotech. For each of those, we have to drive a whole-of-government approach that's able to essentially ensure that we lead in these domains, that Americans are adopting and utilizing these technologies, and most importantly, that we are outpacing our adversaries that are very much on a mission to try to catch up with us.

Essentially, what we have to do as a country is pursue a two-part strategy: one of promote and one of protect. What do I mean by that? If we want to lead the world, or the US as we do as a country want to lead in AI, we have to do both those things. We have to make sure that we are building the greatest technologies here at home and adopting them, and also protecting those technologies from nefarious adversaries that want to leverage them to their advantage. We can begin on the promote side. You can think of promote as kind of three things you have to do and get right. The first is on research and development. The US government spends about $150 to $170 billion a year on R&D. We need to direct that and prioritize those dollars into the areas that are most important for our national security mission, and our sense is it's AI and quantum and so on. So how we spend that money in creative and smart ways so we can keep having the next great breakthroughs is a lot of what our office does and helps coordinate. The second big part of the promote strategy is around deregulation, or the idea of removing barriers to innovation. President Trump has made it very clear that no matter what we're doing, we cannot allow bad regulations to stop us from actually achieving what we want to do as a country. So if we want commercial drones flying for delivery, if we want supersonic jets flying again, if we want autonomous vehicles on the road, if we want AI-powered medical diagnostics approved by the FDA, these are all actions that we need to take as a country to accelerate sort of the removal of these unnecessary barriers and allow this technology to succeed.

The third part of the promote agenda is all about adoption, both domestically and internationally. Take AI, for example. We lead the world for the totality of the stack—from the chips themselves, the Nvidia chips that all of you probably know very well, to the models, to the actual applications. We as a country need to be having our great industry companies, academic institutions, and everyday Americans using this technology. But even more importantly, we need to have the rest of the world running on an AI stack that is ours, that's American. There's no reason why our adversaries should be the ones that are pushing their AI stack around the world. So that's the promote side. Essentially, we have to accelerate as much as possible to make sure that we're making the greatest breakthroughs here in the US. At the same time, we can't be giving away that technology to our adversary and essentially giving them tools that allow them to accelerate themselves. What does that mean? We have to have simple and strict export controls. For key technologies here in the US that, if we protect them, will limit the ability of adversaries to grow and expand, we need to be smart about doing that and make it very clear to industry what the rules are, and to all of our international partners and allies that want to access them. As you know, the hottest topic today in this export control regime is around high-end semiconductors. The second is around research security. If we as a country are creating these great breakthroughs in labs at universities and even at some of our great companies, we have to make sure that they're not being stolen by an adversary that's made it very clear that they want to do this repeatedly over the years. And the last is protecting our critical infrastructure. There is nefarious Chinese technology throughout critical infrastructure here in the US and around the world, and we need to be much smarter as a country to make sure we root that out and protect this critical infrastructure. So zooming back out, to achieve this technological dominance the President is talking about, we have to both promote and protect equally. We can't do one without the other.

Yeah, I think one of the biggest challenges of the previous administration was there was this intense fascination with export controls, and it wasn't just to limit the access of the PRC to these high-end chips. It also became very difficult for our partners and allies to access it themselves, and that's a huge problem. We as Americans have the best technology in the world. We need our friends and our allies to be using it, to be running on the AI standards that we build here in the United States, to be running on our chips, to be using our models. What the Chinese have made it clear is when they're able to actually produce as many semiconductors as they want, they will certainly be sharing those Huawei semiconductors with their DeepSeek model to everyone in the world, if not even giving them away. So, we as a country need to be much smarter, and that's what we're going to be doing in this administration: making sure that our partners and our allies are using our great technology that everybody already wants to use. So, that's why I think we have a huge opportunity.

I think we need to accelerate adoption. I spend a lot of time talking to innovators, to people in labs, to researchers, to scientists, and I'm very confident that as a country, we will continue to make unbelievable breakthroughs in all the technologies talked about earlier. But those breakthroughs are only really valuable if we actually adopt them and allow the American people to fully realize the benefits of those technologies. When it comes to things like AI, we may have the best frontier model that the world has ever seen. It may be released in a couple months and it may be top of the leaderboard on all the leaderboards online, but if no one is using it, if the Department of Defense isn't actually adopting it and using it in its stack, if all of our great financial institutions aren't attempting to leverage those models to drive better services for their customers, it really doesn't matter. So, we as a government are very, very excited about helping industry and helping everyone else in the country adopt this technology. That ties back to a lot of the deregulation issues we were talking about: how do we make it easier for Americans and American companies to adopt?

Yeah, as exciting as the adoption question is, it's also a real challenge. How do we get a lot of these technologies out into the hands, in a very safe way, of American customers, American citizens, and American companies? So we think a lot at the White House about creating an ecosystem of trust where these technologies can be absorbed. The other thing I think a lot about is what our adversaries are trying to do. This has come up a little in this conversation already. There has been a plan for many years for the PRC to try to catch up with the US on these emerging technologies by the end of the decade, and that's something that we in the US need to think about, but at the same time not be obsessed with. We know that we have the best technologists here in the US. We know that we have the best companies. We know that we have the best technology. What we need to be willing to do is actually adopt it, export it, and continue to accelerate that effort. So that's what I think a lot about.