NPF China DeepSeek Webinar Transcript, March 20, 2025
Anne Godlasky, NPF:
Hello everyone and welcome. I am Anne Godlasky, president of the National Press Foundation, speaking to you from the Evelyn Y. Davis Studios in Washington, D.C. Thank you for joining us for “From DeepSeek to BYD, Is China pulling ahead in innovation?”, a free webinar sponsored by RELX, a global provider of analytics tools, including LexisNexis. In recent years, China has shifted from technological imitator to innovator showing rapid cross sector advances, even as American leaders from Biden to Trump. Promise the US will remain the gold standard here to help journalists and others tuning in to make sense of the global innovation race are some incredible experts. So thank you for joining us today. We have Susan Ariel Aronson, a research professor of international affairs and a public interest technology scholar at George Washington University. Aaronson is also co-principal investigator with the NSF NIST Institute for Trustworthy AI in Law and Society and is a senior fellow of economics at the Center for International Governance Innovation.
We have Brittany Nguyen who covers AI for courts and has previously reported for Forbes and Business Insider. We have Horan Omar, who is a senior manager focusing on AI policy at the Information Technology and Innovations Foundation Center for Data Innovation. Previously she worked as a technology and risk management consultant and as a crypto economist in Europe. Marco Richter is the senior director of intellectual property analytics and strategy for LexisNexis IP Solutions and he was the first employee of Patent Site, a tool that analyzes and visualizes massive IP data sets, which we’ll get a peek at later. And we have Professor Christopher Tang, senior Associate Dean of Global Initiatives and faculty director of the Center for Global Management at the UCLA Anderson School. He’s also a thought leader on supply chains. So as you can see for those tuning in, we are in good hands today. So for everyone tuning in, please remember you can submit questions for these panelists at any time by using the chat or the q and a feature and we’ll get to as many of your questions as we can throughout the next hour. So let’s get things started. I’m going to start with AI even though I know that we’re going to cover a lot of technological ground today, but it seems that it wasn’t that long ago that we had headlines calling DeepSeek a Sputnik moment, and then just last week we had headlines asking if Manus is China’s next DeepSeek moment. So hold on, I’m going to go to you first. Could you kind of put this in perspective for us? How big a deal is this both DeepSeek and now Manus and the generative AI landscape?
Hodan Omaar, ITIF:
Yeah, thanks Anne for having me. I think both are really important moments in China’s generative AI trajectory, but they mark distinct strategies and strengths. Deep Seeq was primarily a model driven breakthrough, really showcasing China’s ability to do novel cutting edge AI research and turn that into developing frontier level AI models that can stand alongside offerings from open AI and Anthropic. Manus is slightly different. It spotlights product execution that’s focused on integrating existing AI capabilities, including those from the west into a user-friendly Chinese AI agent. So the real distinction I think is between developing the new and sort of deploying the existing when it comes to AI. And both are really important from a competitiveness standpoint because ultimately for competitiveness, the goal is widely adopted, high impact products and those are two levers to get there. And I think what each of these examples show us is that China is doing both and China is good at both and I think that’s why it’s sort of important but distinct.
Anne Godlasky, NPF:
And we can, I know there’s been debate over whether it should have taken us by surprise or not, but it does seem to have taken a lot of people by surprise. Professor Tang. I’m wondering if you could talk about that given the export controls and the various limits that China has encountered, how did it get to this place where it seems to have made such leaps and bounds and surprise the rest of the world?
Christopher Tang, UCLA:
Thank you for having me. I think that we had to take a step back. I think for a long time. I think the western world always thought of China cannot innovate. I think that is a myth that need to be changed. I think that actually China is very good in innovations. Before they did not innovate much because the markets was closed. There’s no demand. So therefore then there’s no need to innovate because there’s nothing to sell. But now I think that because the world markets, China is pushing very big on innovations. So innovation is leveraged what is already existed and then to make use of it. So in this particular case, because of the export control that US has imposed, so China does not have the access to the most advanced AI chips like the A 100 from Nvidia or the most advanced equipments to produce their own chips from A SML from the Netherlands.
So because of that, they need to work around on what they have, so therefore they use a more second tier level kind of chips like the eight hundreds from Nvidia and then they had to use a different approach to develop the reasoning model for the Dipe. So to compete with Open AI, the o3. So that is really a spotlight moment. It’s a very really innovative way to do frugal innovation. You don’t have everything, but you just use what you have to make something good out of it. So I think that was a surprise. That means that the west did not think about China can make use of what they have to create something good. So that is really an exciting moment,
Anne Godlasky, NPF:
Sort of a necessity is the mother of invention situation. So obviously I know there were a lot of factors going into this, Marco, I wonder if you could talk to us a little bit about the data that you have seen that has signaled China moving in this direction. And I know you have some slides for us as well, correct? Thank you.
Marco Richter, LexisNexis:
Thank you for asking. And actually perfect segue first of all, because you had said it’s a bit of a surprise that we’re all surprised about this moment coming, right? And that we had thought about China not being an innovator. It’s quite interesting because when you look at the actual data, the facts, right, the facts that prove the innovation, the patent data of the world, you see pretty much the opposite. And you also see the rise of Chinese in innovativeness. So we’ve recently went into our data and have pulled analysis on all of the Worldwide’s patents. To put things into perspective here for the audience, there’s about 16 million active patents out there. So 16 million unique inventions that organizations research institutes, companies alike, individual inventors can utilize, enforce and monetize. And there’s about 70 million in total including everything that has expired already. And what you’re seeing on the screen now up first is the share of innovation coming from the different countries.
And you see here over the past 10 years how now China has been outpacing the rest of the world rising to an astonishing 54% of all active patents worldwide originating from China, meaning carrying a Chinese inventor address. So just think also of the human capital and the human ingenuity behind these patents that was built up by China all of these years clearly outpacing most other economies here in the world. Now this was a very willful decision by the Chinese government of the past 20 years to chase and invest into the patent system to drive a huge number of patent filings. And what you will often hear is that China’s chasing numbers but not necessarily quality. So that’s where we’re bringing in a different layer and perspective here. This is the impact of innovations measured through citations and market coverage in a metric that we call the patents index.
And you can see even by that metric, China has been rising quite a bit now, covering 43% of all worldwide pet strength. In other words, they have very relevant innovation. So that patent race didn’t just produce mass, it also produced class. You can also see that the US on the flip side has lost a little bit in share, but it’s punching above its weight When you compare it with just the size of the portfolios, we’re now going a level deeper and we’re looking at the dependency of countries on each other or on itself. And first we’re going to look at inventions made in the United States and how they have been dependent as measured through citations on Chinese inventions. In other words where US inventors thought this is an interesting idea, I’m going to develop this further or going to solve the problem in a different way.
And you can see by the year 2000, this was affecting about 2% of the innovative strength of the United States. That has massively increased by the year 2020 to an astonishing 47%. So in other words, 47% of the entire innovative strength of the US as direct and relevant prior art in China, meaning there is innovation in China that is further spurring also innovation in the United States and where you can infer a technological dependency. So from China not being a player on the world stage in terms of innovation over 20 years ago, this has drastically changed. On the flip side, we can observe around China’s ability to self sustain. So to truly innovate and not just copy a similar effect, while in the year 2000, China was only building for 22% of its patent strength on its own innovation meaning with a lot of the rest was building on prior art developed in the rest of the world.
This is massively increased to now 88%. So truly becoming self-sustaining and driving innovation in and for itself becoming less dependent on other nations. And this is not just affecting the relationship between China and the U.S. This is also true for many other countries in their relation to China, where China is starting to set the pace. And we are seeing that also down on the technology levels. And I’ve also brought one example here for you. One of the hot debated technologies of our times is around 5G interconnectedness of devices, fast and rapid exchange of data and information fueled by 5G and being a hot topic also in political debate. And what we can see here is that China over the past 10 years has really leapfrogged in terms of patent numbers. You see the active portfolio of companies being headquartered in these countries on the x axis while maintaining a high average quality of about three, which represents three times the worldwide average. You also see significant growth in the United States and in South Korea in terms of numbers. However, in the case of the United States that came at the cost of decreasing average quality. And we can only suspect that we will see very similar developments for areas like 5G in the future when we stay with the interconnectivity theme. So I would say if we look at patent data, it is not astonishing that we have these Sputnik moments, right? China has been becoming an innovator and it has laid the foundation to continue to do so.
Anne Godlasky, NPF:
Yeah, certainly it looks predictive when you see it that way. Of course we know as you pointed out, quantity and quality aren’t always correlated, but there are other things you can look into from a trust perspective. I know that that’s a big concern for consumers no matter where they are in the world. Professor Aronson, this is an area that you’ve written quite a lot about. I’m wondering from the perspective of AI governance and trustworthiness, is there a global leader? Is that even the way that we should be thinking about it? I wonder if you could speak a little bit to what we’re seeing right now.
Susan Aaronson, George Washington University:
Sure. Trust is a leap of faith, right? And trust is hard to establish, but it is generally established through predictability, reliability. And so you have a technology that is changing so rapidly and yet so many countries have decided it is first a public good, which means that it is something that they can use to solve, excuse me, to mitigate problems that transcend borders generations. It’s a general purpose technology, so it’s essential to economic growth and it’s a dual use technology, which means it’s also essential to national security. And those three perspectives are contradictory because of the technology keeps changing so rapidly, it’s hard to establish trust. Frankly, there are only two governments that have directly regulated the technology in an attempt to build trust. Those two nations are eu, which is not a nation and China. And I would say that the United States, I’ll just be really brief and so it’d be interesting to hear what my colleagues have to say, but the United States is not establishing trust in that.
Since the switch to the Trump administration, the Trump administration has moved away from a focus on trustworthy AI trying to signal to users that they can trust the technology, that the technology will not be damaging to them and their children or to their jobs or to their future, but to competition in AI. And I think that’s deeply worrisome because we are not going to be successful as competitors in AI if the technology is not trusted, the technology will not be trusted if it is seen as discriminatory or unsafe, et cetera. Thanks for hearing me out.
Anne Godlasky, NPF:
Thank you. And I do want to get back to US policy and regulation, but first I want to make sure that we hear from Brittany covering tech. Obviously you’re dealing with hype all the time and a lot of excitement around new products. How do you put developments like Deepsea or Manus into context for your audience?
Britney Nguyen, Quartz:
Thank you for having me. I would say I have to do a lot of my own research. I think joining panels like this and learning from experts and professors who really get to focus on this technology and the competition with China is really helpful. A lot of my coverage has to be very easy to digest because I’m trying to reach a large audience and I have a responsibility to a lot of people to make sure I’m reporting accurately and able to make things simple. And so I read a lot of research papers from think tanks such as Rand and CSIS. And then again I try to talk to professors and researchers who really get to focus a lot on the competition without influence of focusing on stock market movements or fluff from companies that say, oh, we have the best, they’re not going to catch up to us.
Things like that. I think reading actual research papers and reading even the papers that come out with models such as DeepSeek is really helpful because yeah, you’re right, I think a lot of companies can release something and they just want the most consumer interest to sell their product or to get share from another company that they’re competing with. So I think it’s always important for me at least to lean on people who really have the knowledge and the focus who are cutting through the noise anyways with the work that they do. And I think Professor Tang brought up a really good point that a lot of what we see is Western media, Western focused. Why are we surprised? Because all we do is talk about our own companies and our own innovations. And so for me, I don’t speak Mandarin, I can’t read Cantonese or Mandarin, but I think it’s important for me too to see what outlets in other countries, especially in China and Southeast Asia that are not so, that are not so Western centric, what they’re saying and what they’ve been reporting on. Because otherwise that’s why we’re all caught by surprise because we haven’t been paying attention. So that’s what I think is really important.
Anne Godlasky, NPF:
Thank you. We have a question from someone, a participant who’s tuned in Brianna Noble asks, China’s automotive sector has grown substantially. What role does access to mobility play in China’s ability to innovate? Professor Tang, would you like to take that one?
Christopher Tang, UCLA:
Why not? Well, thank you. Well, I think that we have to go back in time. When it first started, China did not have enough cars. The first car company after the economic reform was a Volkswagen. It was a joint venture with SAIC, which is state owned enterprise. So at the time the Chinese consumers really need a car for mobility. So that’s why China was the biggest car market in the world at the time because the economy is booming and they need transportation, they need to get things moving. So that was very good. But then when the market so big, everyone Russia in, so GM and Mercedes-Benz, everyone was in that then. But China pivots, they realize that they cannot compete on the internal combustion engines. They’re too late to the market. So therefore 20 years ago they pivot into electric vehicles. They said, well this one Germany has not quite developed yet, and yet China is making a lot of batteries.
Batteries for cell phones, batteries for other things. And Japan was actually letting go of some of the batteries technology. And then China start developing it, as Marco pointed out then said, well, once we deliver batteries, why not? We want the battery for cars, for drones. So the next wave is want to develop the batteries for drones, and that’s why after drones, why not cars? So that’s how it’s evolved over time. So now actually they are leading the markets in the world in terms of EV batteries and electric vehicles. So this is really astonish shaped development in terms of r and d patterns that Marco points out. And also from the research to development to commercialization. So they were very good in pushing for the cars now is actually they’re building up factories in other countries. Now the flow is reversed because now China see that this Europe, they’re behind in development EVs. So well that’s not, why don’t we sell the cars? And now why don’t we produce a car in Europe? So that’s why BYD is now having a factories in Hungary to avoid the terrace and then the next one could be Turkey. And now they’re thinking, well, why don’t we make the car in Germany? So this is amazing. That is a lake commerce now become a first mover in Europe.
Marco Richter, LexisNexis:
And if I may, Christopher, right, all these technologies that you mentioned, right? When we look into the data, we see these. It’s not a surprise. The innovation data shows it. It shows it years ahead of time. And then we in the western world, we look astonished like, oh, it’s becoming reality guys, it’s not a surprise. It’s happening in wind power. Now China is starting to produce windmills here in Europe. European wind power producers for decades felt like, oh, it’s going to be too expensive to ship it over here, right? Guess what? China decides to build them here and now they’re all getting insecure. So
Christopher Tang, UCLA:
Yeah, I would like to add one point, if I may. Mark brought up a very interesting point. I think the west, even my students, they get confused between invention versus innovation. I think for the western world, they say, well, we’ve already done our cars. The cars become monopoly, it’s also commodity. So in the case we focus on pure science. So in terms of pure science, I think US is the leading. So for example, during covid us, it was really the only country can develop the mRNA covid vaccines. China could not do it. They try the chip manufacturing, the design, the advanced technology for the equipment is done by the Netherlands, the S-A-S-M-L, which is derived from Phillips originally. And then in the U.S. we have a MD, we have Nvidia to dev all this, but China focused on applications. They said, okay, you can do your pure science.
We are going to read your papers. That’s why, that’s how the citations got reversed At first we’ll learn from you in terms of what kind of inventions you have, what kind of new ideas you have, and then we leverage that to develop something useful, something apply something I can make money out of it. So they’re very good in apply research. That’s point number one. Point number two is that the sheer size of people, the talent, they have equivalent number, same number of pure scientists between US and China, but they’re five times more engineers. So that means that they can convert all these new ideas, all these patterns into products now because they’re also manufactured in China. So the learning this feedback loop, the learning curve is much faster than the west because in the west we don’t produce too many things. If you don’t produce any things, your learning cycle is much longer because you need to produce something before you learn something. But in China, they can learning by doing the learning very fast. This is exactly what happened to BIDI took my students to BID factories 10 years ago. They laughed at it because compared to the German student from BMW said, this is a joke. But 10 years later, my God, now you see what happened. They already introduced DeepSeek into their car. We don’t even have the basic AI in the car. So actually in the sense we are behind,
Anne Godlasky, NPF:
Susan, I want to give you a chance to sound off on this, and
Susan Aaronson, George Washington University:
Dr. Tang made some excellent points. I’m really grateful in terms of diffusion, though the United States does have excellent infrastructure for translating research into actual business use, but that’s not the only factor. And I worry very much about what’s happening in the United States, predictability, reliability, and when you’re trying to govern a system such as AI, it’s really important that the government clearly signal to market act towards what it’s going to do and how it’s going to do it. And the United States has moved quite far away from what it was doing just three months ago. And so it’s that certainty and stability that market actors need, especially as the technology evolves rapidly and such as AI or XR or any other technology that’s built on data and essentially is opaque to users. So coming full circle in terms of trust, I do think that the Chinese government with its regulations on data, on data as an asset on AI, is actually at an advantage because it’s created a market environment where innovation can happen, but also where it’s clear where the government is going to play a role. Same with the eu. And I think this is really going to be a problem for the United States. Be keen to hear what others have to say.
Anne Godlasky, NPF:
Thank you. Hodan, I wonder if you could comment on some of what Professor Aaronson and Professor Tang mentioned about how this learning is happening in China?
Hodan Omaar, ITIF:
Yeah, I think one of the things that Professor Tang mentioned that I was thinking about is as we talk about how many companies there are, we’re talking about DeepSeek and Manus, but when it comes to generative AI startups, China has more than 250. And before we were talking about DeepSeek, there was already an understanding. I think that there are leading a group of leading unicorns in China. These include GPU AI, Moonshot, MiniMax, 01.AI, Byron AI, right? These five unicorns between them proved that they are doing innovative research simply they’re doing things their own way and they certainly have a vision for themselves. A lot of them are actually focused on AGI. This is not simply a Western vision for artificial general intelligence. There are several Chinese players in the ecosystem that are also reaching for this kind of moonshot project. AGI is not some simple productization, it’s a vision for breakthrough change, the game type of AI.
So yeah, just on a point that I think what’s interesting about those five companies though they’re all in some way tied to Tsinghua University, whether they came out of it, whether they spun out of it, whether they are faculty there, whether they are researchers there, they all have direct ties. Can you imagine if open AI Anthropic scale AI and Databricks all came from Stanford, the way that it seems that this one university is really a backbone for the kind of Chinese generative AI ecosystem, I think is really interesting. And so understanding the ways in which that learning that Professor Tang happens and how does innovation spill over, we think of Silicon Valley in China. It seems even more kind of small. And I think that’s super interesting. And then to Professor Aronson’s points, I also think the idea, I certainly agree that the United States needs to come up with a vision for AI governance that is clear and that, yeah, when it comes on the global stage that it can champion because at the moment it doesn’t have that. And I think for this conversation, I think that’s really important because it’s important for its ability to influence the global conversation about this. Thinking about the African Union and Africa as a continent, which is an incredibly important geopolitical partner to both the United States and China. China has come to the African Union, they have signed partnerships.
The African Union has commended and shown that they will align with China’s approach to AI governance. It’s because the United States is also not really presenting a credible offering that could be an alternative. And yes, the African Union in and of themselves and the continent as a whole, it’s not a monolith. Kenya is doing its own thing. South Africa is doing its own thing. Nigeria is doing its own thing, but in each of these places you can really see the ways in which China has been a credible partner for a long time. And so moving to AI, it also makes sense and that’s something that I think US policy makers need to wake up on.
Anne Godlasky, NPF:
I do want to continue on US policy, but quickly could you just share why those five companies you mentioned, why did you call them unicorns?
Hodan Omaar, ITIF:
Because I think the definition of a unicorn is if you make over a billion dollars or something. So they’ve passed some threshold in terms of the amount of investment that they received.
Anne Godlasky, NPF:
And I forget, I apologize, I forget who made this point about, I believe Professor Tang, it was you about the mRNA vaccines. And Marco, I know from the LexisNexis top 100 innovative companies in the world, the US still dominates in terms of pharmaceuticals and medical technology specifically. But when you’re looking at quantum computing, which I know I do not know much about, but I know that China is making inroads there that could impact all those other things, national security but also medical technologies. Marco, I don’t know if you want to speak to that or if any of the other panelists have any thoughts on that.
Marco Richter, LexisNexis:
I think quantum computing, like many of these top technology topics like battery semiconductor in general, et cetera, we basically see the same trends and patterns everywhere is rapidly innovating in these spaces. And ultimately, yes, innovation is what the consumer pays and what generates revenues, right? I’m also want to be mindful I of that. And so we see a race between the countries and the research institutes and the corporates in the respective countries to have technological leadership also in quantum computing and in terms of quantity, we already see a ton of output there from China. Well, in terms of quality, the US still has an edge, but I think it will take a very concerted effort to maintain that. And we’ve seen the releases over the last couple of weeks about new quantum computers being launched, that war of some of what were considered the most advanced ones so far. And ultimately I think what it comes down to for quantum computing, one of the other panelists said it earlier is then going to be the application layer. What are we going to use these massive F forces and computing powers for, right? Like being at the next breakthrough pharmaceutical truck applications, be it medical devices, MRT, screening technologies, etc. Right? It really will come down to the application then.
Anne Godlasky, NPF:
Thank you. Hodan? And then I have a policy question for Susan, and in the meantime, please attendees, anyone who has any questions, feel free to put them in the chat or the q and a and we will get to those as well.
Hodan Omaar, ITIF:
I just wanted to quickly chime in on the quantum aspect. Quantum is a sort of umbrella technology. So you’ve got quantum computing, quantum communications, quantum sensing, and it’s interesting to see including with patents, where is China leading? Where is the US leading, where are other countries leading? And interestingly, the United States still has a relative lead in quantum computing in terms of those kind of breakthroughs, but when it comes to quantum sensing, which has much more national security implications, that’s actually where China seems to be doing really, really well and also putting a lot of its emphasis. And so I think breaking the technologies down into the different subparts and then thinking about what the implications of that particular one is because quantum sensing for instance doesn’t, while it has highly high national security implications, the economic implications are relatively small. And so that’s how I tend to think about those different technologies.
Christopher Tang, UCLA:
Yeah, I would like to add to this point as well, I think that’s both Marco and Holden has really made a very good point, but I think that in terms of the gap between US and China in quantum computing is closing Marco’s, right? US is still leading, but the gap is not that. Why also bear in mind right now, I think that in terms of the quantum chips designed, Google is leading, they have a sycamore chips, but in China they, Baidu and Alibaba, most people did not think of Alibaba as leading. They are leading in AI as well as the quantum computing. Now in China, as I mentioned earlier, they’re very good in applications Now actually we need to think ahead of time, what do they do with quantum computing so they can actually use it for AI? Because right now AI require a lot of computation power.
That’s where quantum computing can play a very important role. Now, how to integrate these two technologies, we don’t know yet, but it is in the world, so we have to keep an eye out there. Then the question is that how would quantum computing AI combine, what do they use it for? Now in China, it’s leading in one thing, undeniable facial recognition. They are the leader of the world. Now, if they can combine these two, my god, the facial recognitions will go beyond that. Also, they can incorporate this kind of quantum computing AI chips in class in the future that would actually could really reach in terms of truly autonomous driving. And also they can use AI to synchronize all the traffics along the road because the idea of a smart city, they try to push for the next wave of smart city such that our traffic light will synchronize with the car movement because sometimes they find it so stupid, there are no cars on the road. But then the red light, so what is this? But in the future with AI, with quantum computing, they can actually have autonomous really streamlined traffic congestions in many smart cities. In China, for example, sun is already pushing for this kind of smart city with autonomous, with synchronous traffic flow, which is really fantastic.
Anne Godlasky, NPF:
Yeah, bringing it back to the human level of second, you mentioned, you mentioned facial recognition, which brings up privacy issues between that, and obviously we’ve seen this with social media as well. When you have the technological landscape evolving so rapidly, we have seen policymakers in the US as well as elsewhere, struggle to adapt to that. Susan, I wonder if you could weigh in on what policymakers could be doing to respond more quickly, better particularly in the us.
Susan Aaronson, George Washington University:
Well, first of all, I do want to say this because I think it’s very, very important. AI is a global product. It’s built on global expertise, it’s built on global data, it’s built on people trained from around the world. It requires huge amounts of energy in the current scaling models, et cetera. So we need to think about it in terms of any regulation that is domestic. Unless it’s a huge country, it’s really hard to regulate this technology domestically. That’s the first point I want to make. But since the United States and China are the leading venues of AI development, I think it’s important that those two nations play a leading role in building trust. What is regulation? Regulation is a tool, a commitment advice that is designed to build trust so that people who are affected by AI have recourse or employers that use AI. So governance needs to be transparent, it needs to be accountable, it needs to be effective.
But when a technology is changing rapidly and when governance is captured by the very firms, and I will say this, I mean my own research looks at public comment on AI around the world, and I am really shocked to see how unresponsive governments are to public concerns. In general, the public doesn’t comment, but most of the people who do comment are already heard because they’re tend to be industry or industry associations. So my big long-winded point is that governance needs to be democratically determined, it needs to be transparent, it needs to be accountable. Right now in the United States, we don’t have that I would say to some extent is because government is captured by very rich people who have huge investments in AI and want to see AI used for national security purposes and want to see AI widely without constraints as it develops. And I don’t think that’s good for American AI because I want to see AI reach its potential.
And to me that potential is not only as the general purpose technology, but as a public good. For example, everybody cites ENTs think tank every week comes out with a report on how AI is being used in, for example, the sciences or commercially. And this technology could do so much good for the world and is at times, but unfortunately, to achieve that objective, it needs to be governed. I’m not saying regulated and you can regulate in lots of ways. Sorry for the long-winded answer. You can regulate the business practices, you can regulate the technology, you can regulate the use, you can regulate by type of AI. We are not brain dead, but you have to regulate it systemically and that’s really hard to do because it is a system.
Anne Godlasky, NPF:
Do you envision a global body to regulate or enforce around AGI similar to around nuclear weapons? I mean
Susan Aaronson, George Washington University:
I think that is not going to happen. Nuclear weapons were developed by governments, really by scientists, I should have said, but they were funded by governments in this case, yes, there’s government funding, but there’s lots of private sector funding and the people developing AI tend to work for corporations, not universities. So it’s not a comparable situation. The risks are different, the rewards are different, et cetera. So I don’t like that analogy at all. But the United States is a signatory to treaty language already related to AI by the council of Europe, and it’s not perfect, but it’s a start. I believe if we had, and I’ll be explicit, more thoughtful policy makers, we would not be doing what we’re doing on AI. We, but the United States comparative advantage doesn’t last forever, and competitiveness trust is a key factor in competitiveness. So I’ve come full circle here, but eager to hear what the others have to say.
Anne Godlasky, NPF:
Going back to BYD for a second, just because I know that’s another hot area. Obviously it has come out saying that they are three to five years ahead of everyone else in terms of their technologies. Marco, I wonder, I know that China also has NIO, which is looking at swappable batteries for EVs and that sort of thing. In terms of is that thing if journalists look at patents or ip, is that something that they could better get ahead of what the next big features are going to be?
Marco Richter, LexisNexis:
I would say when non-patent people look at patent patents, what the advantage that they’re getting is they see where companies are going in the future and if they’re putting their money where their mouth is, you bring up BYD as an example, and we’ve seen, I think was it early this week, late last week, they shocked the Tesla stock by announcing a new fast charging system where I think within five minutes you can charge about 300 miles of range, which has so far been unheard of. And turning then attention to the patents, BYD has been investing in charging capabilities and the respective batteries for many, many years and is dwarfing, for example, a Tesla in terms of top patents, they have about five to six times more top patents in their portfolio in these areas than a Tesla who’s often being looked at as not being innovative. Ever since model three, model Y were launched, right? You have some cosmetic, but the technology has really been outdated, so it’s no surprise that Chinese competitors are passing Tesla left and right on our audubons here.
Anne Godlasky, NPF:
Britney, I want to stick with speaking to our journalists and the audience. You obviously have been covering this for a long time and these companies, while some talk about being open, they obviously can be tightlipped as well. How do you go about building sources in this space?
Britney Nguyen, Quartz:
I think it’s a good question. I think honestly, LinkedIn and finding people that work at these companies takes a lot of trust and building trust with people if you’re going to tell their stories. And I think again, just talking to researchers and people in the space who are not involved in open AI or Google for example, who are working in startups and competing with these companies, I think they give a lot of insight into how models are being trained and built because they’re doing it themselves and again, they’re competing and have a lot more insight than surface level that I would have. So I think just talking to knowledgeable competitors has been really helpful for me to take a step back from open AI’s reasoning models for example, what does that even mean? And I think when Deeps seeq shocked the stock market, that’s most of what my coverage was about, apart from trying to explain what deeps seek is and sort of what it proved, I remember just talking with some startup founders and they were like, well, I don’t really believe that they only use this many GPUs. I think all these, they’ve been stockpiling, high flyer capital has been stockpiling and I don’t really believe this research paper and here’s why. And so I think that has been helpful for me to look with more critical lens at what larger companies are saying and releasing.
Anne Godlasky, NPF:
Thank you. Hodan, I wonder if we could go for a second. China obviously is able to give quite a lot of state support to these companies, but Biden and Trump both have issued multiple executive orders on AI. Vance told Europe just recently that the US is and will remain the gold standard. What steps is the US taking to basically do that and do you think that they are sufficient or will prove to be?
Hodan Omaar, ITIF:
Yeah, so I kind of think of the government’s role in two buckets, at least for my own work and purposes. The first is really around getting the governance right like Professor Aronson was talking about. And then the other bucket is the proactive role for government in ensuring that AI achieves its potential, like closing this gap between what the potential is and what the reality on the ground. And I do think that there is an important role for government there because it’s not going to happen on its own, especially for those areas in which I’m seeing comments around smart cities. Smart cities are a public good. There’s no company that’s by itself going to invest in ensuring cities are as capable as they can be as well built as they can be, as innovative as they can be for citizens and for the average American. The adoption of AI in healthcare ensuring that it’s done correct, these sorts of sectors, education, healthcare, the city, these are particularly areas where the market incentives are not to adopt AI optimally. And so that’s where the government has a real role to play. I think that recently we’ve had the AI action plan requests for information, and so it’s clear that the administration is going to have an action plan that they come out with soon, and I think what I’d love to see in that is a proactive ensuring that there’s this kind of proactive role for government in addition to having that voice on governance that’s sort of credible around the world.
Anne Godlasky, NPF:
Thank you. Professor Tang, did you have anything you wanted to add there?
Christopher Tang, UCLA:
Yes, I think that’s Professor Eon point of a very good point in terms of we need to have a systemic governance, but it’s very difficult to do because I have a panel discussion with lawyers and also with the people from the DC because in terms of AI, we still don’t know exactly how it worked. Is it still a black box then when you don’t have black, you’re doing black box, how do you have regulations and governance? It’s very difficult to do. So I think that right now, I think this is similar to social media, they can do a lot of good things AI, but they also have bad actors do bad things. But then the question, how do we regulate those is regulating the outcome, the actions, the systems, the products, which level are you going to govern? That has not been subtle now, but given that then we still need to compete in terms of AI development, that’s where China comes in for the data.
They say, OK, we need to protect ourself, but in terms of development, they’re willing to try different things. Now that is the alignment between the government companies, scientists and consumers. So that’s why China allow and push for autonomous driving. They learn, of course they learn and correct mistakes, but that’s why they can push out the autonomous driving in terms of technology faster. But then on other hand, in the western world, we worry about privacy issues, safety issues and all this. Where do we draw the line? If you want to compete, we have to take risk and move fast. If we want to really be very safe, check all the cross, all the T, all the i’s, yes, we safer, but we behind the curve. Then the question, where do we draw the line? We don’t know. The second point I’d like to make is that Marco and you partner BYD, besides BYD, there’s another company that the reporters may want to look into because talk about Brittany tried to look in terms of new, this company is called Sami.
OMI is actually starting with there other leaders in smart homes, but they just introduced the ev. Now how the linked to market, how do the compete are beside the design is beautiful, but that is not it is they allow the car to actually integrate with your smart home devices such that you can actually manage your home maybe then beyond maybe your AC before you get your home, your social system, you can all the food such that there’s someone who deliver the dinner to your home by the time you get home to synchronize it. So I think that is the next excitement for the new consumers. That means that you could integrate your entire lifestyle inside the car with a smart home devices integration. So that’s something the reporters have not really looked into this really the emerging EV company to look into is sell me. Thank you.
Anne Godlasky, NPF:
Brittany, do you want to speak on that?
Britney Nguyen, Quartz:
On sell Me or Yeah, I mean I think that’s a good point. Again, there’s so much focus on whatever’s moving the market. So DeepSeek for example, I mean their v3 model came out in December and then R1 was happened to be the same day as Trump’s inauguration, but I think they had released it a week before and then just the market crashed and everyone was like, oh, it’s because DeepSeek said its models are comparable to open AI and Anthropic. It’s so much cheaper. And that’s why Nvidia lost almost 600 billion in market value and like Professor Tang was saying, there’s so much innovation happening right now that the west is not covering or we’re not as focused on until it impacts us and our investments and our companies and how we see ourselves in competition. I think another good example was going back to export controls, like when Huawei released the eight 60 and apparently the chip was really impressive and the US kind of was shocked by it, but then I think Gina Raimondo ended up saying, oh, it’s not actually that impressive. But I think there needs to be focused by the media and by developers that we’re not the only ones here. There’s other competitors. It’s also not just China. And I think there just needs to be more dialogue on how we compete and how we work together. And it’s not all just fighting, it’s also collaboration.
Anne Godlasky, NPF:
On that note, I know that we’re getting near the end, so I want to pose this question to all of the panelists because I’m very curious to hear your thoughts. So looking out over the next year to three years, what issue related to innovation obviously, do you hope that journalists are going to really dig into what major technology should journalists be watching particularly closely, especially in terms of how it will impact the citizenry, the average, the public, the average consumer? So let’s go with who wants to take it first
Susan Aaronson, George Washington University:
I’ve got to run. So I think I’m going to go.
Anne Godlasky, NPF:
OK, great.
Susan Aaronson, George Washington University:
OK, so two things that I may, digital twins, which to me are one of the best technologies ever. You can model the earth, you can model a flood zone, et cetera. So I hope they look into that. But the other one is the opacity, the black box nature of AI, and I’m done. OK, thank you. Thank you.
Marco Richter, LexisNexis:
I can go next. I would ask journalists to leverage data because I think data is an unbiased source of information. It also solves the issue that we have all here in the panel addressed that we are over-emphasizing what we’re seeing in our quote media, what we’re focused on by what’s driving our government or local governments, et cetera. If you turn to data, you get an unbiased story. You can look for the signals that are out there and you can identify the areas that are trending and the companies, the nation states, the technologies that are setting these trends. So my call here for the audience, for the journalists is leverage the data that is out there. In particular, the data around inventions, innovations, and patents.
Anne Godlasky, NPF:
Thank you. And for the journalists who are on the call, Marco’s colleague, his contact information is in the chat if you would like access to that information that he showcased earlier. Professor Tang.
Christopher Tang, UCLA:
OK, thank you. Well, I fully agree, Marco. Data is the most unbiased source of information, and LexisNexis is fantastic. So I think in terms of topic, I’m writing op-ed for the tariffs issues and the impact on global supply chains because I think that’s what President Trump is doing now is going to reshape the global supply chain configurations. Now, this one, you can use the shipping data and also some of the data available to actually understand how the impacts in terms, how it changed the landscape. That’s the first topic. The second topic, I think that that is really big for the world is the AI in healthcare services and drug development. Because right now I think in the us, our our GDP, the bulk goes to healthcare. Then the question is how will AI to actually make our healthcare services and development drug development become more efficient and also more cost effective? So I think some of the data could be very useful. I think that is something we should keep an eye out. Thank you.
Anne Godlasky, NPF:
Thank you. Hodan?
Hodan Omaar, ITIF:
I think I would say bringing it back to the beginning of this conversation, what are other ecosystems doing? The piece I really liked from last year was the financial Times. This piece that spoke about four of the companies that came before DeepSeek, and that really helped shape how I thought about that DeepSeek moment. Is it really a surprise? No, the FT was talking about those companies before. What’s happening in Europe, ra, what’s happening in the uk? What does the AI ecosystems or any innovative ecosystem, what does it look like? Who are the players? Who are the people? I really also enjoy, I don’t think that from the ThinkTech space, it’s hard to tell the human stories behind some of those, and I think that’s something that journalists are well poised to do.
Anne Godlasky, NPF:
Never go wrong with focusing on context. Thank you. And Britney, fittingly, you have the last word here.
Britney Nguyen, Quartz:
Thank you. I think what I want to focus on and what I think is getting coverage but deserves a lot of coverage is the environmental impacts of all this technology that we’re developing, especially here in the us. The call for millions and dollars towards data centers. I mean, I think there needs to be a lot of looking into how it’s impacting the environments they’re being built in because they’re being developed, they’re being built in more rural areas and it’s going to impact wildlife. It’s going to impact the quality of living for smaller communities that live around data centers. There’s noise pollution, there’s all these issues that are going to touch people that don’t have any hand in AI development. And I also, another thing is autonomous weapons is something that I think is getting a lot of funding. There’s AI companies starting to really rev up work with the US defense and intelligence agencies now under Trump.
I think there’s more comfort between Silicon Valley and the Pentagon and policy makers. And so unfortunately, a lot of AI development is about the money and it’s about getting as many contracts as you can, as much product out there. But there’s people on the other end of these products that are going to suffer even if they have, again, nothing to do with AI models. And so I think a lot of coverage is always on chatbots. It’s on what? What’s the most consumer facing AI product or AI tool? AI agents, everyone’s talking about AI agents now, but there’s other ways that AI is being implemented. I think Professor Tang brought up healthcare and there’s a lot of positives and as Professor Aronson said, but there are ways that it’s harmful that isn’t just a chatbot that’s gone awry. It’s in everything. And I think there should be more focus on the harmful impacts that are just not as obvious.
Anne Godlasky, NPF:
It’s certainly very ripe for both “new you can use” kind of stories as well as deep investigations, and I’m sure that we will be seeing a lot of those over the next year or so. We are over our time. I want to thank everyone who joined us today, especially all of you panelists. It’s been a fascinating conversation, so thank you for sharing your expertise. Thank you for everyone who tuned in and thank you to the webinar sponsors, RELX and LexisNexis. For the journalists who are on the call, if you have additional questions and would like or would like to reach out to any of our panelists for an interview or a story, please contact Alyssa Black at ablack@nationalpress.org. For those of you who are new to NPF, the National Press Foundation’s mission is to make good journalists better through free training, and we announce that in our newsletter. The link to that is in the chat, so please sign up for those notifications. We currently have two fellowships, which are longer trainings, which are expenses paid, and those are open, so please check those out. Thank you again to everyone and happy reporting. Take care. Thank you.
###
