Shownotes Transcript
This is CNA Talks, the flagship podcast of CNA, a Washington, D.C.-based research and analysis organization. Technological innovation is a critical element of strategic competition between the United States and China. And in the 21st century, the line between civilian and military technology has become blurred.
In this episode, we examine developments in China's AI, subsea sensing, and commercial space industries, and the intersections between their civilian and military purposes. Welcome to CNA Talks. I'm John Stimson.
Today, we are going to discuss the latest issue of Intersections, CNA's news digest describing the interplay between the People's Republic of China's, or PRC's, technology acquisitions and defense industrial-based development efforts, U.S. response, and emerging technology risk with military implications. And I'm excited to welcome two of the authors of Intersections, Chris Cairns and April Herlevy, to the show today to discuss this. Chris and April, thanks so much for being here.
- Great to be here. - Thank you. Great to be here, thank you. - So let's get right into it. With everything going on, I wanted to start with the tech sector. Let's talk about some developments in that area. I understand from your report that the PLA's Academy of Military Science developed their own LLM called Chat Pit. Can you tell me a little bit about that?
As you mentioned, ChatBit is an LLM or large language model, which is similar to other large language models that folks may be familiar with, such as ChatGPT is the best known one. And it's a question answering tool that, at least according to this Reuters report, that the PLA is using for intelligence analysis. So presumably they're feeding intelligence related questions into it and having it crunch through the intelligence data that they have at their disposal to come out with
answers, whether it's analytic answers or summaries or whatever it is. And the interesting angle of this story that we covered is that they didn't invent Chatbot from scratch. They're actually building upon an existing publicly available LLM developed by Meta called Llama. So Llama is Meta's widely publicized LLM. It's
similar in structure and usage to ChatGPT. And the important characteristic is it's freely downloadable by anyone. So you can access it anywhere in the world without any restrictions.
I think that's really interesting. And I was I noticed this when I was going through your report is it sort of reminded me of how like some of the technology and spy craft happened in the Cold War, where because the United States has an open system versus the Soviet Union's closed system, the Soviet Union had a clear advantage in the spy war. I know it's not an exactly one to one parallel here because Meta is a private company, not a
government entity. But, you know, I do feel like that interplay still exists between the two. Yeah, I think so. And it kind of represents a fundamental tension in the development of LLMs, which is that for companies like Meta and others, the reason that
they fund these what are called open source software efforts in general, not just for LLMs, but for all sorts of things, is sort of a rising tide lifts all boats. The idea is that there's a collective benefit, which will directly benefit the company, but also just benefit the industry in general to putting out these sort of software frameworks that people can download and then modify and build upon. And that's been true for a long time. So Silicon Valley is
for LLM development is thriving off of the basis of just making the software freely available. But of course, that means that other countries with other potential strategic objectives can also access the software and adopt it to their own ends that might not be consistent with US national security. April, anything you want to add on this point?
The only thing I want to add is sort of the larger big picture of what the People's Liberation Army, we're in a phase right now where the PLA is moving from what they used to call informatization to intelligentization. I hate these words because they kind of don't translate into English well, but it's this idea in terms of
intelligence gathering, they do really want to move to using these sort of large language models and artificial intelligence. And that's like what they call intelligent warfare. So I think this story just really shows that this is the direction they're moving and they are trying to figure out how to use open source and publicly available information, just as Chris just beautifully outlined, but use it to achieve their military end state. So I think just understanding it in the larger context is really important.
Got it. Thanks for that, April. And yeah, the big picture look at this is great. And kind of, you know, in a related note and sticking with AI a bit, I understand from the newsletter that a PRC semiconductor company is attempting to expand into the chips needed for artificial intelligence. Can we talk about that a little bit?
Sure. So the company is SMIC or Semiconductor Manufacturing International Corporation. And this is China's largest chip manufacturer. It is designed to be a direct competitor over the long run to the better known TSMC, which is, of course, Taiwan Semiconductor Manufacturing Corporation. And the PRC government is committed
basically giving large amounts of subsidies to SMIC. There's tremendous startup costs associated with getting all the infrastructure in place, much of which China has historically had to import, such as chip fabrication machines, which they have imported until recent restrictions from the Netherlands, from ASML is another company that produces those machines.
So the PRC government is fronting a lot of money toward this end. And it raises the question, which we also raised in this story, how financially viable is this? Because despite all the advances that PRC companies like SSMIC have made in recent years in chip manufacturing, their yield rate, which is to say their hash
how low their defect rate is in churning these chips off the assembly line is still significantly higher than, say, TSMC in Taiwan. And so basically, it's unclear whether they're actually able to make money or whether the only reason they're able to continue production of these chips is through the PRC government subsidies. It reminds me of like a surge we're seeing in nationalism, just sort of
generally right now. And I wonder how much of like that kind of an approach is as much about like, you know, signaling towards building things within China as it is about like the, you know, the sheer need for the actual production of it. I don't know. I guess I guess think that concept of like when something is being done for a PR perspective is interesting as well as it being, you know, like a strategic approach.
I think it's all of the above. So certainly there's a PR aspect to China having a national champion that it can point to proudly as we now have a world-leading chipmaker as well. But Xi Jinping also, so that's a PRC top leader and CCP leader.
Chinese Communist Party General Secretary Xi Jinping also views having a thriving domestic semiconductor industry as a strategic imperative. He's mentioned before how semiconductors are to industry and to the economy like human heart is to the human body. Basically, the idea being we can't survive and we can't function without it. So he wants China to be the world leader and to be self-sufficient in semiconductors.
Now, is that easy to achieve? No, because currently semiconductor supply chains are global. They're enormously complex. They involve thousands of suppliers. And it's very, very difficult for any one country to go it alone in practice. But to your point, I think not just China, but also the United States and other countries are recognizing for national security, maybe we want to invest to have as much of the industry within our own borders as possible.
Yeah, and just to build on what Chris just said, I mean, I think this is a trend we're seeing in a lot of countries, this intersection between national security and economics. I mean, one of the reasons we chose to call our newsletter Intersections is because we were on the front edge of seeing this trend play out in practice. We've seen it in the United States. We've definitely seen it in China. We've also looked at partner and ally countries in previous issues of our newsletter, such as Australia or the European Union.
And so there is this recognition that things that used to just be in the economic realm now really do have national security implications. And some of our older frameworks that were developed in the Cold War to think about these concepts don't always work as well. Like the idea of dual use technology, it's hard to determine what's dual use when so many of the technologies we now use could be used for civilian or military purposes and often both.
Yeah, and I think this next section that we're about to discuss covers that really well. The featured section of intersections in this issue was PRC advances in subsea sensing. You know, Chris, I'm thinking that not everyone might be familiar with the concept of subsea sensing. So can you give us a little background and what that is?
So subsea sensing is vital to both the naval and military domain, but also to civilian oceanographic research. And there are many different technologies involved in subsea sensing. So you think of, for example, acoustic sensing, so picking up sounds below the water, which of course is important for sonar. You think of magnetic sensing, so there are actually sensors that can attempt to detect
magnetic field anomalies that are caused by large metal objects moving below the surface of the water. So you think of using that to try to detect submarines. There's technologies used for seabed mapping, which is important in the civilian domain for seabed resource exploration, for example, so deep sea mining, but also obviously naval and military domain for anti-submarine warfare to know what is the underwater topography. So it's a very different field with a whole lot of capabilities
technologies involved. And so in intersections, you reported on some developments in this area, primarily the Highway GD 11000. Can you talk a little bit about what that is and how it plays into this idea of civil-military fusion that we've been discussing? The Heidi GD 11000 is a deep-sea cable-laying ship. And what makes it special is it's
basically that it's able to lay cables at greater depths than the PRC has previously had the capability to do. We're talking about winch systems that can lower a cable and, say, sensors attached to the cable into the ocean at great depths. They're very heavy. They're difficult to work with. But
The Highway GD 11000 has an improved winch system that is able to handle longer and heavier cables. And as a result, they can now put sensors down into areas that were previously accessible, such as, for example, the Mariana Trench, which is located in the Western Pacific.
You think of Challenger Deep, and you may recall James Cameron descending down to the bottom of Challenger Deep in a televised special some years ago as sort of a stunt to show how exciting it is to do human exploration of such great depths. But the Mariana Trench is also important for military reasons because it's located very close to
to what's called the second island chain in the Western Pacific, which is where U.S. military bases, notably Guam and also on the Northern Mariana Islands are located. It's a key naval stretch of territory of water as well. Does this achievement surpass what the United States is capable of doing? Or is it more like China is now able to do something that the United States was already doing?
I want to give a little bit of the background context because for much of the last century, basically American, European and Japanese firms were pretty much dominant and sort of working in the underwater space. And what's important to note, though, is that basically since the early 2000s, China has recognized that they were far behind in this domain. And so basically the Communist Party in the early 2000s set out to develop ocean technology
technology as one of its core missions. And they specifically in a lot of those early works talked about they need to do more surveying and mapping work. So things that the US and Europe had been doing for a long time, they weren't doing as much of that. And then more recently in China's 14th five-year plan, they actually explicitly said
they want to develop "real-time online monitoring systems and overseas observation stations for the marine environment." So some of these technological upgrades and the things that Chris is talking about, in the past, the Chinese state was not able to do this stuff. There's now companies, there's now technologies, and they're building out this capability. And so, like I said, this has been a goal of theirs for a while, but that's because
in earlier periods of history, they just weren't as focused on working in these environments. Got it. Thank you. And so that kind of gets back to maybe what we were talking about a bit before of this global wave towards a little bit more of like insular nationalist approaches to these things. Whereas in the past, countries may have relied on, you know, like the global economy and specialization. Now there's a pressure and a need to be self-reliant and more so than there may have been in the past. Yeah, I think that's right.
It's really interesting to look at the PRC entities, and by entities I mean both research institutions as well as companies that are to various degrees ostensibly private, meaning they're not necessarily obviously state-affiliated, but in practice many of them have ties to the Communist Party or to the PLA. And there's a lot of sort of fluidity and overlap
in terms of China's civilian oceanographic research with some of these same entities also supporting the PLA. So in the example of the entities involved in the Highway GD 11,000 program, we have Dalian Maritime University, which as we note in intersections has previously done research for the PLA's
Central Military Commission. So that's the top leadership body chaired by Xi Jinping that oversees the PLA. And then another entity, which again is at least in how it represents itself as a private firm, Hengtong Huahai, supported the SHIP project
but has also been flagged by the U.S. government as supporting China's military modernization, which makes sense going back to my earlier point about how often a lot of the same technologies, such as acoustic sensing, that are useful for civilian oceanographic research also have naval or military applications. Finally, I want to turn to the commercial space sector. And I understand that the PRC has undertaken several commercial space and low-earth orbit projects with the goal of competing with companies like SpaceX.
Can you tell me about these and why they're important to the PRC? I think it's interesting to note that China is investing heavily in both their military and now their commercial space sectors. And I bring this up because historically, or at least over the last couple of decades, it was really more on the military side that they were focusing. But because of the Russia Ukraine war and the emphasis that's been placed on Starlink,
the leadership in China has realized, "Oh, wow, this is really an emerging area that's really important." And space innovation is explicitly part of what China calls its strategic emerging industries.
And there was a comment in 2023 by Chairman Xi Jinping where he basically said, we must not fall behind in this area. What I think is one of the most interesting developments on the commercial space sector in China, though, is it does seem to be happening from the bottom up. There's clearly been top down guidance. So the central government has said that we want to focus on this. But right now, there's a lot of basically municipal developments.
and provincial authorities that are encouraging this kind of development. And so specifically, the example we talk about in this intersections newsletter is the Shanghai Space Com Satellite Technology Company, or its acronym is SSST. And that company is backed by the Shanghai municipal government, and it's raising funds to launch a commercial satellite constellation known as ThousandSails.
The stated goal of that program is 600 satellites by 2025 and 15,000 satellites by 2030. So that's a really aggressive program given that they've only recently started these efforts. And what's also interesting is this company is already out negotiating with other countries about bringing high speed internet services to those countries. For instance, there's been an agreement recently signed with Brazil and that opportunity presented itself partially because Brazil
was in negotiations with Starlink, but those ended up falling apart. So this is an area where the Chinese government sees both commercial applications, just building out companies that have these capabilities, but that could eventually be used for military purposes, as we've seen in Russia, Ukraine. Chris, anything you'd like to add on this point? You know, I hesitate to point the finger too directly that the evolution of the PRC's commercial space sector is...
in some way mirror imaging the development of the US commercial space sector. It's very easy for us to look at what China does through the prism of our own commercial defense industrial base when that's not always the case. But I do think certainly the example of SpaceX and Starlink and how successful that's been getting up and running and expanding is definitely a model there for the PRC to look at.
As more of these projects come online, as we see like low Earth space and there's more of these private companies launching satellites, you know, I can't help but worry about the risk of space junk. And that's like, you know, some of these satellites colliding while in orbit or colliding with an out with a decommissioned satellite. Is there anything that these firms, you know, Chinese, American, whatever, to cooperate to limit those risks a bit?
I think there's probably more that can be done, but I do think you raise a good point that as more and more of these constellations are put up in the sky, I do think space debris in general is going to be an issue, especially debris that might fall back to the Earth. I mean, for the most part, that's usually not a danger specifically, but there is more and more discussion about space.
what are the implications of all this space debris? And one of the things I'll just mention that it's not ready for publication yet, but there has been some other work within our division on the space domain. So everyone should keep in mind that we have a few more studies coming out on that topic in the next few months. So keep your eyes on the website for those developments in the space domain. Thanks, April. All right, we're just about out of time, but I did want to give you both a chance to give your final thoughts and anything we might have missed in this discussion. April, we can go to you first for this.
As I said already, with the issues about dual-use technology, I think there's so many fascinating developments going on where it's really hard to differentiate between what a civilian use and a military use is. So I really encourage people to read the Intersections newsletter so they can get a handle on this. And I think that's a discussion that we're going to have to have as we move forward in the future, thinking about how the U.S. deals with technology policy. Chris, what about you?
I would just point again to this sort of inherent tension that we see again and again, and we've covered this in a number of different intersection stories and previous issues. This tension between innovation, which on the one hand, in many cases involves international research collaboration or collaboration between companies, particularly in areas like LLM and AI development, and innovation
Conversely, this drive for national security and sort of owning aspects of the process and owning technologies within national boundaries to make sure they serve one's own national security and not necessarily adversarial objectives. This is not going away and it applies to many different technology areas that we see. Absolutely. And all the more reason for people to check out intersections. Chris and April, I want to thank you so much for coming on the show today. Thank you so much for having us. Thank you very much.
For our listeners, there'll be a link to the latest edition of Intersections in the show notes. And from that page, you'll be able to explore any past editions. There's a lot of great information there. So definitely check that out. And if you'd like to subscribe to Intersections, you can email intersections at cna.org. But I want to thank you all so much for listening. And we'll see you next time on CNA Talks.
The views expressed are those of the commentators and do not necessarily reflect those of CNA or any of its sponsors. CNA Talks is produced, edited, and mixed by John Stimson. Our theme music is by Edward Granga. If you enjoy our show, we'd love it if you could give us a five-star review on Apple Podcasts and tell your friends about us. Thanks again for listening, and we'll see you in two weeks.