From Action Plan to Mission Impact: Allies, Adversaries, and AI

In this episode of INSA's From Action Plan to Mission Impact: Key Takeaways from America's AI Action Plan podcast, host Chitra Sivanandam sits down with Chip Usher, Senior Director for Intelligence at the Special Competitive Studies Project, to discuss how AI is reshaping global competition, diplomacy, and national security. Together, they examine Pillar Three of America’s AI Action Plan, focusing on international leadership, trusted partnerships, and strategic competition with adversaries such as China. Chip shares insights on AI standards, supply chain security, global adoption, and the importance of building trusted AI ecosystems with allies. They also discuss the role of public-private partnerships, AI security, and how the United States can advance a democratic vision for AI while strengthening its position in the global technology landscape. 

Transcript

Chitra 0:03

Welcome to InSA's All Source podcast from action plan to mission impact. I'm Chitra Sivanandham. Yevgeny isn't able to join us today, so I'll be leading the conversation. Uh today's episode, Allies, Adversaries, and AI, focuses on pillar three of America's AI action plan, which puts international leadership, partnership, and strategic competition at the center of U.S. AI policy. AI is not anymore just a technological race. It's a contest of influence, trust, and national power. The action plan makes clear that the US works with our allies, sets standards, and counters adversary advances to shape the global AI ecosystem for years to come. In this conversation, we're going to explore how the U.S. and its partners can build trusted AI alternatives, strengthen our alliances, and protect critical technologies while staying ahead of emerging risks. Joining us today is Chip Usher, Senior Director for Intelligence at the Special Competitive Studies Project, who brings a front review on how AI is reshaping strategic competition. All right, let's get into it, Chip.

SPEAKER_01 1:17

Thank you, Chitra. Great to be here.

Chitra 1:19

Awesome. And for those of you who don't know, Chip knows everything about intelligence, national security, and our lovely elephant.

SPEAKER_01 1:27

Oh, pardon me on failure, Chitra.

Chitra 1:29

Okay, so the action plan kind of framing AI as a matter of global influence, security and trust, as we were mentioning. Um, we're really talking about international AI diplomacy as a high priority for the administration. How do we do this, right? What is at stake for the US and how do we lead shaping this global AI adoption standards, governance, cooperation?

SPEAKER_02 1:52

Well, first of all, thanks for having me. It's great to be here. The partnership uh between our two organizations has been great. Um at uh the Special Competitive Studies Project, you know, we have long focused on the power, the transformative power of artificial intelligence uh for our economy, for national security, uh really for our way of life. And the uh uh the pillar that we're gonna focus on today is something that we have long agreed deserves focus. And and the reasons are kind of multiple, but uh, artificial intelligence really does have the potential to upend like the global order, right? What will matter for national security uh in the decades to come will include a nation's AI stack. That will become an important factor. And we're seeing you know, whispers whispers of this or hints of this even today in the conflicts in Iran, where Iran took it upon itself to strike three data centers owned by AWS and Bahrain and the United Arab Emirates, sort of indicative of how they evaluate the importance of AI infrastructure. And um, you know, the the other point is that AI, unlike other technologies that we've been dealing with in recent years, you know, really is going to become an integral part of the infrastructure.

Chitra 3:21

Aaron Ross Powell And probably more than just public infrastructure, just like everybody's day-to-day infrastructure. Trevor Burrus, Jr.

SPEAKER_02 3:26

Right. So not just a product, but we're building critical infrastructure here that didn't exist just a few years ago. So with that frame in mind, uh setting the standards is incredibly important because this is what is going to sort of define the parameters for AI adoption in in the years ahead. And this is an area that uh rightfully the current administration with their with their plan have identified as a priority. And it's certainly an area that our main competitor in this arena, the People's Republic of China, are already paying attention to. So, you know, with that, uh I I mean I think that in and of itself speaks to the importance.

Chitra 4:10

Trevor Burrus, Jr.: Yeah, and and I think it's it's important, right? Because I think everything like lives or dies on interoperability. So standards is the fundamental component there. I feel like in um in the past, both in the US kind of uh development of standards and even international standards, I feel like standards generally lag the market. So with AI moving as fast as it does, and with adversarial activity creeping up on us like before we can even like figure anything out, like how how do we really do this? Like it seems like an impossible task.

SPEAKER_02 4:44

Uh it it is challenging. Uh and the first challenge is you know, standards are are boring. Yes. Uh people are much more interested in in developing the capability and and and rightfully so. And I would say, you know, the current administration, while it's it's you know, added this to the pillar and identified uh competing in standard making bodies or setting standards as an important objective, uh, the primary emphasis has been on let's let's build it and let's let's accelerate the move in that direction. Um but you know, China has you know shown itself rather adept in recent years and other standard making bodies like the ITU or IEE, uh where they they get their voices heard and we're sort of stepping back from some of these international organizations and they're stepping into to field of interesting.

Chitra 5:34

That seems like there's uh an adversarial risk there.

SPEAKER_02 5:36

Absolutely. Absolutely. And so the you had uh you had asked, so so how do we go about it? And and you know, it's sort of defining with partnership with with industry, and now that uh set of industries are gonna be the uh foundation model developers, the data center uh uh builders, uh all the companies involved in the so-called AI stack, uh, to figure out what uh AI, uh American AI values and standards are to look like and then push for those.

Chitra 6:07

So that's an interesting thing. And like even the way you are describing it, right? American AI be curious to figure out like how do we know what that really looks like because there's so much you know dependence and cooperation, and this is not just like globally, like in some cases, maybe even with our adversaries, right? So what what is American AI versus non-American AI? And we're how do we know when we've like you know pat crossed that line?

SPEAKER_02 6:34

Aaron Powell I'm not sure we're when we'll know it's it's happened, but I think it's being built right now. So we're seeing it sort of in a in a nascent stage. You know, some basic things that we're we're seeing are uh just kind of the comparison with uh trying to deep seek or quinn or you know, these other things, or the way that um uh guardrails and safety um uh uh is infused in uh American developed AI systems by choice, by the producers themselves.

SPEAKER_01 7:05

Based on ethics.

SPEAKER_02 7:06

Based on ethics that, you know, in the case of some of our AI companies, you know, trace right back to the American Constitution, right? So these values are are sort of baked in and they're reflected in the outputs that uh our our AI systems generate versus what you get from something from China, where uh you know they don't have as uh much concern for human rights or minority rights uh or accurate portrayals of history and that sort of thing.

Chitra 7:33

Yeah, I think that's uh a very good point because we had a conversation yesterday about like um testing and bias and things like that. And um I think what you're describing is like on one level you could think of it as bias, but it's really like us encoding those morals and values in our systems to be more fair and and open, but it it is kind of reflecting the way we see things and the way our allies see things. So I think that's an important distinction. So as we think about um, I guess like US diplomacy in that frame, you know, what does that look like?

SPEAKER_02 8:13

Well, I'm I think it's gonna be critically important for the United States to put forward a positive vision and a positive, I'll I'll use the word offensive, not just a defensive plan, but an offensive plan. And by by that I mean it's not just enough to say, you know, stop using Quawe, don't use Quinn or don't use DeepSeek, bad, bad. Um there are reasons to have those concerns, and we we ought to be upfront about that. But we need to kind of quickly follow with um sort of an integrated, coordinated, um, positive vision to um third parties, uh third country markets with what an American vision for AI looks like. I mean, this is sort of what underpins the State Department's pack silica uh campaign, where um I think it's up to seven or nine countries have have joined with the United States in seeking to build the silicon AI infrastructure of of the future. Um I think that's a good first step, but it reflects um, you know, the desire to get kind of lead from the front foot. And so what that's gotta include is a vision. And then um the full tech stack, meaning the algorithms, the data, the data centers, the pipes and networks that go along with what uh other parties are looking for for their for their AIs. And I would argue there there needs to be at least awareness or some sensitivity to what other countries, other societies are going to seek from AI to preserve their history, their culture, their way of the Peter Peter.

Chitra 9:57

Yeah, no, I like that. And I think that like ties in really nicely with like this pillar three notion of like the deepened collaboration in order to get to more trusted systems. And, you know, our allies need to trust what we're providing. And it seems like it's this tie between not just the software environment and um those capabilities, but also the supply chain. Um I think that's what you're you're hitting on. So how I feel like that's where it gets murky, right? And that's where um there could be like erosion of trust. How do we how do we better um manage that and as we build more capabilities, like work better with our frustrated partners and our allies?

SPEAKER_02 10:37

You're absolutely right. That the the trust factor for for everybody, right? Any any any customer now who's thinking about using clawed code, you know, is going to be immediately confronted with, you know, do you give it access to your data and to your files? And uh AI systems that uh demonstrate good custodianship of uh customer data need to be uh celebrated, uh, need to be promoted within the American AI vision. And you know, instances where you know uh AI developers are uh mishandling, misusing, misdirecting data, right, that needs to be called out. And it should be uh you know something that we're unafraid or unabashed about doing. Um it's gonna it's gonna take time. Um, there's sort of innate distrust in Europe and other places of Silicon Valley. Uh they are far behind the United States and China in their development of their own sovereign AI capabilities. So that's it needs to um uh you know, there needs to be patience and there needs to be uh time, but there needs to be a focus on it.

Chitra 11:53

Yeah. And I think that gives us a better global ecosystem with our uh partners. I think that also touches on the action plan that has this kind of coordinated approach where it proposes more heightened cooperation through organizations such as State Department, USTDA, Exxum, DFC. Like so what are your thoughts on what that type of like ecosystem play looks like as it relates to the full stack delivery?

SPEAKER_02 12:19

At SESP, we've sort of promoted the idea that there needs to be uh you know a technology czar, there needs to be a uh a technology ambassador uh who is out uh taking point, whose main job is to build this coalition and bring to bear sort of the full range of capabilities of the US government to help sort of lead this vision. And yeah, you know, I we've talked previously about an American vision for AI. It starts there, but you know, ultimately I think it's gonna be sort of a democracy vision for AI. Um and uh the United States has a number of tools in its tool belt to provide financing, help establish um uh uh standards and um provide assistance and that sort of thing. But you know, you know well the US government is uh main challenge in life is coordinating with itself. And so having a point of focus probably in the White House, uh we think is going to be the uh initial first step. And then uh, you know, engaging allies and partners, sort of maybe building outward from close close friends and allies and building outward from there, perhaps build based on Pax silica is uh is the way to go.

Chitra 13:38

Yeah, I really like that approach. Um I hadn't thought about that, but we're it seems like we're already on that journey, right? Even the existence of these chief AI officers and all these agencies. Um it all developed out of this necessity to say, hey, there's so much happening so fast we have to uh better coordinate amongst ourselves. And I think that's likely true with everybody else that we're working with, right? So it seems appropriate if we had kind of that technology czar that has a uh a a different optic, right? It's not an economic play, it's not it's it's uh you know us building that almost the same thing, like that, you know, global industrial base on AI.

SPEAKER_02 14:15

Absolut absolutely right. And this and this person could be sort of the quarterback when uh we're engaging with, say, African countries, where they are you know interested in developing uh and enhancing their AI capabilities for their own national goals, so for their own economies. And the United States can step forward, uh maybe under the the stewardship of the direction of this of this AI czar to kind of put the play together about how do we coordinate uh between the uh Development Finance Corporation, the Department of State, Department of Commerce, and um a uh select number of of U.S. companies who can provide the capabilities and the guidance uh that our partners uh would need. And you know, that I think is entir entirely within the realm of the possible today. And it and it's you know, there's need to do it.

Chitra 15:12

Yeah, no, I agree. And then I I would hope that we would have um comparable thinking with our allies so that we can actually do this at that global scale, right?

SPEAKER_02 15:22

Aaron Powell And in a in my experience, um the the way to sort of bring people on board is to actually build stuff.

SPEAKER_01 15:31

Yeah.

SPEAKER_02 15:31

Right. And I'll I'll take uh the I I come from the intelligence community, I spent 30 years at CIA. Uh we have invested quite a bit and have a very uh robust set of strong intelligence liaison partnerships uh all around the globe.

Chitra 15:47

Yep.

SPEAKER_02 15:48

Uh in the future, those intelligence liaison partnerships need to necessarily include some element of AI. And what might that look like? Right we right now, when we share information with our friends and allies, oftentimes, I'll be honest with you, Chitra, it's it's 19th century technology. It's here here's your intelligence, here's your intelligence. Maybe a phone call. Don't photocopy it, right? Um but uh down the line, we need to be you know cooperating with uh, again, on a on a selective basis at first, but sort of building outward, you know, AI-driven platforms for this sort of information sharing and insight creation. Yeah, right. Um we can be tackling regional or global challenges in partnership with our AI, um, our intelligence allies in an AI-empowered environment. And that will kind of bring them naturally closer to us technologically, but also politically, diplomatically. So that's uh an example of how the power is sort of building things.

Chitra 17:00

Maybe it's like the uh the agentic framework on diplomacy, right? Where it's like, hey, we finally don't live in our little silos doing our own things and saying information sharing is the delivery mechanism. It's it's more how we work back and forth and collaborate and cooperate, right?

SPEAKER_02 17:16

Like that you know, it can start with shared shared awareness. Yeah. Um, an event has happened, uh, and everyone's you know keen to get the the intelligence reporting from the United States to explain what happened. But as we developed um, you know, collaborative AI environments, we can be working side by side with allies to enhance our mutual insight into these developments where they're sort of they're participating, therefore persuading themselves and seeing things our way much more quickly and effectively.

Chitra 17:50

Yeah. So that kind of makes me think go into the supply chain discussion again. So if I think from the contractor world perspective, like, you know, it's easy to just think about like one company, but really like the strength of your company is the strength of your like partners and teammates and everything. So it's kind of like this layered effect. Same way if we think of our AI, you know, infrastructure and our people and our technology and layer that with all of our allies, that's a very complicated supply chain. And I know that one of the other things that the plan highlights is, and you we talked about it earlier with like semiconductors and the silicon um aspects. Like, where are the big the big risks, right? If we really think about this like very stringy layered um thing that exists, like how do we how do we like really close those gaps properly and uh respond and make sure that we we know where our risks and vulnerabilities are and what's what's preventing us from doing the things we have to do on that?

SPEAKER_02 18:46

Well, I think I think government's gonna have to play uh a key role. I I have a sort of a two-part answer to this. Um government's gotta play uh a key role in first sort of establishing the requirements, okay. Um, for Pax silica participation for an an American-led diplomatic effort around AI, there needs to be some guidance provided about what is and what is not going to be acceptable with regard to supply chains involving Chinese or other um adversarial state tech in that stack. Okay. So that's step step one. Um the the the the the risk that I think we all need to kind of bear in mind is that adversarial risk, because while we're doing our thing, China, uh whose AIs more frequently are based on open source models are poised to be adopted at a much faster rate and much more broadly than American AI, which is mostly, with you know the exception of Meta and and some others, um, you know, more on the closed source model system. So um we it it's actually it's it it's becoming apparent in some sort of strange and interesting ways. You know, I've talked to a lot of American tech companies that are very interested in using AR for various mission purposes, including in support of the U.S. government, including in support of the Department of War, the intelligence community. And they're very honest to like, you know, Quinn's pretty good. Uh Deep Seek is pretty good. R1 is is not a bad model. And I can get it pretty cheaply and I can I can get it to do the thing I wanted to do. I get it to do the thing I wanted to do, and I can tailor it and it and I can get it fast. Okay. So I'm I'm gonna sandbox with Deep Seek, and then in hopes of, you know, uh maybe I can find something that an American company can provide. But for our foundation model uh participants, you know, there needs to be sort of an awareness that if you set prices too high, if you set uh barriers too high, uh you may end up losing the adoption race uh to China, which is going to be um, you know, much cheaper, uh maybe not as effective, but maybe not too bad and acceptable enough uh for other uh market participants. So you know, there needs to be this uh emphasis, even by our industry, on the speed to adoption. Because, you know, in a way, these markets in Africa, Latin America, East Asia, uh the Near East, these are the battlegrounds, right? Between our vision of AI and Beijing's kind of feels like the um the cellular network issue all over again.

Chitra 21:45

Right.

SPEAKER_01 21:46

Right, right, right.

Chitra 21:47

Um so how do we how do we prevent that from happening the way it did on like the the network sprawl?

SPEAKER_02 21:53

Um Well, I I I think it's uh again this public-private partnership that has to be Be fostered and a set of priorities, sort of building from what this administration has started last year, carrying it forward this year. And as I say, you know, appointing somebody at the White House who sees it as their job and then starting to build things, the solutions that work for the United States, that work for our allies and partners overseas is probably the next step. So they see the benefits and the differences between our vision of AI and what others are providing.

unknown 22:39

Yeah.

Chitra 22:39

And maybe we have to do something a little more than just like a subsidy model in order to get the market and the adoption that we need if we wanted to go in a certain direction.

SPEAKER_01 22:50

Right. Yeah.

Chitra 22:51

Okay. So the the other thing that the action plan directs us to is um thinking about how we work with like things like Department of Commerce. And it's it it kind of implies then like, do we need to rethink things like ITAR? Like how what is your stance on what we have to do to think about protection of our IP along with like export? And what's what's the happy medium?

SPEAKER_02 23:14

Great, great question. And um, I think we ought to be setting a uh pretty firm line with regard to uh protection of data and protection of IP uh for our uh foreign partners. Uh you know, it's it's gotta be unacceptable uh to uh build something in conjunction with the United States only to have either lax IP controls or purposely share uh with China or competitors same thing for U.S. persons information, et cetera. I think we need to kind of set a firm line. Um but we shouldn't, you know, be setting the bar too high too early in an ITAR-type way that prevents American companies from being able to export as they need to in some of these battleground states for precisely the reason that they are battleground. So if if we um make the perfect the enemy of the good, then we're sort of destined to to lose a race here. So I think we need to accept some risk in service of a long-term goal of you know building the AI ecosystem uh that attracts uh countries in some of these uh these markets uh to the US vision. Um and there will there will be some mess-ups, there's gonna be some leakages, there's gonna be some failures. Um but again, I don't think we should make uh the perfect the enemy of the good.

Chitra 24:54

Yeah, and I think the way you're describing it then, I'm curious your thoughts on like who owns the responsibility in this partnership, right? I feel like traditionally we do these things, and if we want the outcome to go the way we need it, we probably have to put a lot more energy and effort into the partnership, regardless of what our allies put into it in order to make it work. Like how much of this do they own and bear the responsibility for? And how much do we own and and where how do we get to where we want to be, knowing that it's all still resource limited and we're running really fast?

SPEAKER_02 25:28

Well, I you know, it start it starts with engagement and conversation. Um I think it it probably requires a little more push from Washington for um uh uh sort of a diplomatic push, uh government-led push to bring people together around um you know this AI vision. Um and the the responsibilities need to be um discussed and shared and aired sort of from the get-go. But I think for certainly our some of our close close allies and partners, there's there's an understanding and a and a desire for you know cutting a cutting-edge AI capabilities, capabilities that um you know, reflective of democratic values. Um, and there's I think a natural inclination that we can leverage and work with to want to partner with the United States. I mean, that's frequently what you hear everywhere you go uh on technology issues is our preferences to work with an American company. Can you just can you just show up? Can you just be there?

Chitra 26:31

Yeah. I mean, I I think that's exactly it, right? We have to really like amp up that um collaboration and communication. And I think one of the areas that there was a call out in terms of uh you know, focus, and probably you know, COVID probably is still on everybody's mind, but how do we think about things like these like bio threats that are um potentially more real and could be crazier given the advancement of AI? Like where where do we look at the evaluation responsibilities and those cooperation responsibilities as we build these partnerships?

SPEAKER_02 27:03

Aaron Powell Yeah, I'm glad I'm glad you raised the bioterrorism risk because it's it's real and AI um has all sorts of benefits in the realms of biology. Uh however, it does dramatically lower the barrier for malign actors and adversarial states to uh use AI to kind of rush uh their development of a bioweapon. All right so now here in the United States, uh we have thought long and hard about this. There's a National Security Commission on biotech that just issued its report last year that SESP uh uh collaborated with closely. And uh our our companies are building the necessary guardrails uh sort of head off, head off this uh uh being our AI is being used for these purposes. Um but with open source models uh being developed in China and elsewhere, uh we can't be assured uh that they are. And this ought to be a key factor in the sales pitch, right, to uh democratic and friendly peaceful nations is that um you know our uh AI systems sort of come with this baked in, and we're willing to work with you and we want to have the discussions with you about what your standards and the protections that you want to see put in place. Um and I think we should be using this uh on you know the bio uh biotech threat, the bioweapon threat, but on a on a host of others as well. Um you know, AI is um one of the the reasons that it's so important to global security is it's uh a a wonderful platform for malign actors to per uh uh promote their information space, their promote misinformation or disinformation, right? So you know a AI that comes from uh the United States and other democracies should be uh in you know less tolerant of misuse of AIs for those purposes. Um we should be developing the ecosystems that sort of call these things out um when they're when they're being done.

Chitra 29:26

Yeah, and maybe maybe we have to do more than just um be the creators of the great tech, right? Because um it makes me think it it it could feel if you're well if you're really way behind, you could think like, hey, this is how I catch up. And if somebody's offering me not just the models and these fundamental capabilities and the better chipsets and the the data centers, right? Maybe if they also offer me the um the extra expertise and all the other components of the infrastructure to get me to catch up fast, like it's very attractive. And um, I'm wondering whether we are um whether we need to do more actually to get that putting in place.

SPEAKER_02 30:11

Absolutely. And um, you know, that also includes AI security in the full sense of the of the term. Um as these uh systems become more vital to our economy, more um you know, important to our societies, thinking about securing our AI tech stacks is going to be a growing area of focus and concern. And it's you know, it's multi-level. Uh it definitely includes physical security. We're seeing that sort of play out in the Near East today, where some data centers are vulnerable to Iranian attack. Uh they're also vulnerable to insider attacks, they're vulnerable to sabotage, all those sorts of things. But they're also vulnerable to, you know, adversarial AI attacks. Yeah. Others seeking to pollute or um, you know, misuse or abuse uh AIs that are being created, um, the full pantheon of cyber attacks. So equipping our friends and allies around the world with even how to think about it, uh solutions uh being uh delivered by uh the United States that include a security program and a security plan would be very attractive, uh I would think, to uh various countries around the world. And it should be part of the conversation.

Chitra 31:39

Aaron Ross Powell Yeah, and maybe maybe there's actually a hidden um story here on the resiliency too, right? Because I think you saw if you s talk about like the the UE data centers, I mean, yeah, it was kind of irritating for a little bit, but everything kind of bounced back, um, taught us something. But maybe uh we have a better story with our partners on how we can offer better resiliency compared to our adversaries.

SPEAKER_02 32:07

Yep, absolutely that's part of it. And um sort of interdependency um uh rather than just dependency. So, you know, I think uh history has shown that when when China you know provides you know 5G solutions or other tech solutions now to include AI solutions, you know, there's always uh in in the deal that China gets access to all your data.

Chitra 32:32

Yeah, you made the deal with the mob. Right.

SPEAKER_02 32:34

You made a deal with the mob. Uh and yes, you got some capability, but you know, it can be turned off like that if Beijing so decides. And because of China's national security laws, you know, all the data that your data center, your systems are generating is is shipped back to to Beijing for their companies to use to compete against you or their government to use for its purposes. And with the United States and our allies, I think we should be, you know, engaging allies and partners in an honest conversation about how they can protect their data, uh, what the what the bounds of their sovereign AI mean, um, and that you know, we present a more positive vision where allies are cooperating, helping each other. You know, that that can include uh you know data that maybe resides uh in a in another region or another area, uh, but is protected and that they have control over.

Chitra 33:31

You know, it makes me think like, um, so you guys in SESP are like a unique organization and you get to have the benefit of working on the national security sector, working um understanding like diplomacy and and government and policy, um, and having a little more of that international um optic. Are there comparable organizations in our allied partners that we can interface with or that are helping to lead that story and work not just government to government, but across a commercial industry that supports the PRO?

SPEAKER_02 34:03

Yeah, there there are some, um you know, less well-developed in in some countries than in others, but uh you know, we're we're finding good partnerships, uh non-governmental partners included in Europe. I'm thinking of the Jedi Organization, uh, there's ORF in India, uh, there are some government sponsors like the UK AI Safety Institute, you know, that are uh eager to uh to learn, they are eager to partner, and they're eager to build. Um and you know, that is also part of this American approach that you know we as um uh a nonprofit organization, we as you know insta-member companies ought to be thinking of participating in, right? So there's what the government will be doing, what the government will be leading with other nations, but there are track one and a half, track two, industry-to-industry collaborations that should be happening in parallel.

Chitra 35:00

Aaron Powell Yeah. It makes me think like maybe maybe there's a piece that's required around more of a uh collaborative international uh international kind of sandbox environment that allows us to do things that we can't otherwise do rapidly, right?

SPEAKER_01 35:15

Aaron Powell Yes. Absolutely.

Chitra 35:17

Aaron Powell Seems like there's some it would be interest, I think, and um it'd be super useful. I think that even across our industry, right, in national security, like working across the little silos has always been a challenge. And I think that um the ability to protect while you collaborate seems like it's vastly improved with today's AI.

SPEAKER_02 35:35

Aaron Powell You know, I'll I'll go back to the uh intelligence example because that's what I know best. But uh, you know, for for years and years, uh CIA and other US intelligence services have trained our foreign partners, right, in how to do analysis. What is good analytic trade graph to create your own PDB for your for your country, right? So with AI, this can be sort of put on steroids and accelerated, uh, where again, they're generating insights for their national security purposes uh with an AI uh a US provided, uh US uh endorsed or uh uh condoned um AI stack, and it works to the benefit of their security and ours.

Chitra 36:20

Yeah. The stack, the thinking and the cooperation, like um all of it. Okay, so as we kind of uh wrap up, I'm thinking like, okay, in order for us to implement this action plan successfully, in order for us to meet all the objectives that this administration is setting out to do and figure out what this global landscape will look like given that it's changing so rapidly under our feet. What do you think things are gonna look like like five to ten years from now? And and what indicators do you think the US and our partners should be paying attention to as we figure out how to make the right steps into the future, like in in the sense of global security and global cooperation?

SPEAKER_02 37:01

Okay. So a a lot of people who uh talk about AI are very focused on the capabilities of the latest AI models. Um how good is is Claude uh 4.6 Opus versus Chat GPT? Um and that is uh a key metric. You know, we're it's something we can measure. Right. And we're you know, we're we're all sort of racing, you know, perhaps to AGI, um, artificial uh general intelligence, you know, this kind of mythical finish line where we've we've built something that is as as smart as a human. Um but there's uh for purposes of this pillar in this discussion, I think there's a more important um index that we should be paying attention to. And that is in five, ten years, when you look at the deployment of AI, um who accounts for most? Right? Um we may have the best models and they may be the most expensive, and they may be mostly concentrated in the United States. Or are our models uh ones that are used all around the world? Or are those Chinese models? And you know, that is a a different set of metrics, a different area of emphasis. And I and I think in five years' time, we want to see uh US uh built, uh or at least US-derived, um, AI algorithms and tech stacks be the standard that most other countries, most other economies uh want to participate in. And I do I do worry that if that is not realized, that uh China would uh sort of uh exploit that advantage to sort of further chip away at uh U.S. global standing and uh our economic competitiveness.

Chitra 39:04

Aaron Powell No, I think you're right. Um makes me think of um social media, and I think for years we were very myopic around our own adoption that we didn't pay attention to the platforms nobody was using here or that were being widely adopted elsewhere, right? And I think it's the same potential to get blinded by our own consumption.

SPEAKER_02 39:23

Aaron Powell Yeah, if we don't watch it, this could go the same way as Huawei, the same way now of electric vehicles. Um, we're sort of myopically to borrow that term, focused on Tesla, which is you know the preeminent electric vehicle manufacturer in the United States, but it's it's not globally. Yeah. Right? Uh it is uh Chinese companies that are becoming the global standard or having the greatest market share globally. And you know, that's a future that we don't want. Uh we want um a uh a US-inspired, US-built, uh democratic, value-ladened uh AI future. That's what we want to build together with our partners and friends.

Chitra 40:07

Yeah, no, I agree. Um if you caught one of the earlier episodes, Mark Muntel mentioned uh we were hoping our AI doesn't end up blowing up the world, right? And so I think you're absolutely right. And uh if we take the positive optimistic outlook, then what would you recommend to like the in-some membership community or industry or a commercial company that does nothing with the government? Like, how do what would you offer them as a piece of advice as we kind of really try our best on this um this model and this pillar with global cooperation and and our our work with our international partners so that we don't get caught like we did with Huawei?

SPEAKER_02 40:45

Okay. So um, you know, the first is to you know think defensively, uh, think defensively about your supply chains, um how uh you know whether you are uh embedding uh Chinese models too too deeply in your strategic plan um and what the long-term consequences of that are. Uh the second is to think offensively uh in terms of uh participating uh with uh the US government in uh Paxilica or other diplomatic initiatives as you're you know engaging with foreign partners, you know, think proactively about uh the other elements of the tech stack uh that those foreign partners may need and build build the partnerships uh as you can with uh American uh uh industry and government to help uh kind of deliver those and participate. Participate in the standards discussion, participate in the diplomatic outreach discussion, uh maybe these track two conversations. Um I would think that would be a great way.

Chitra 41:52

Awesome. Thank you, Chip. Thanks for joining us and for this great discussion. Um I think it's always fun when everybody and that that's exactly why we're here with INSA and doing everything, because we're participating and trying to like set the way um I do feel positive, right? We I think we can we can do this. Um and we can all work together as a community and and get this right, I think.

SPEAKER_02 42:17

Well, absolutely. And thanks to INSA for putting the spotlight on this great, great topic. Uh you're a terrific organization. So thank you.

Chitra 42:24

So in our next episode, um, Yevgeny and I are gonna sit down to wrap things up, reflect on some of the key themes that we've heard from our speakers, um, talk about what it means to really execute on this AI action plan and to deliver real mission impact and do it in a way that um doesn't blow up the world, I guess. Thank you all for listening. I'm Chitra Sivanadam, and this has been INSA's all source podcast from action plan to mission impact.