the courts, campaigns, and more from the world of politics, all at your fingertips. you can stay current with the latest episodes of the washington journal and find scheduling information for c- span tv networks and radio. c-span now is available at the apple store and google play. scan the qr code to download it free today or visit our website, next, the u.s. china economic and security review commission regarding china's military using emerging technologies that impact competition between the united states and china.

investments, artificial intelligence and >> our second panel will address the military applications of chinese investments in artificial intelligence and quantum science. breakthroughs in these fields could lead to a paradigm shift that can lead to a paradigm shift in the way the war is waged with the revocations for regional and global power. we will start with of the senior fellow with the indo pacific security program at the center for the new american security.

mr. stokes previously served in the white house on the national security staff for the vice president then joe biden and a senior advisor to the national security advisor as well as acting special advisor for asia policy. he will provide testimony on battlefield applications as well as broader implications of a and development for u.s. -- rather u.s.-china strategic competition. this is mr. stokes first time testifying before the commission. next, the policy research at the rand corporation. prior to joining rand he was the editor of the china brief at the jamestown foundation and he will examine efforts to leverage models in its approach to cognitive warfare, particularly for manipulation -- social media. he is also a new voice at the

commission. then we will hear from doctor edward parker a physical scientist at rand where his current research focuses on emerging quantum technologies, ai and cybersecurity. prior to joining rand he received his phd in solid- state physics at the university of california santa barbara his testimony will address chinese aspirations to integrate chinese technologies into its military and draw comparisons between the u.s. quantum and chinese quantum industrial basis. this is also the first time for doctor parker testifying before to the commission. your for your testimony. i would like to remind you all all of our witnesses, please keep your remarks to seven minutes to preserve time for questions and answers. i do want to just quickly comment as with our last panel, we are learning a lot. if

doctor parker, you can describe for me quickly later on quantum physics, i would appreciate it. i attended a full disrupted technology course at mit for three days and still don't know. so, i appreciate mr. stokes. >> good morning. thank you chairman and commissioners for inviting me to provide testimony on this critical topic. my presentation will look at the progress of developing military artificial intelligence within the people's liberation army. applications related to an accrued autonomous systems and battlefield functions and support. i will also look at implications for the military and security aspects of the china strategic competition. at the broadest level, china takes a view on military ai potential to help become by mid

century if not sooner. beijing considers military ai to be an essential component to reach the level of military technological development it calls intelligent -- where they supercharge the combat power. in october 2022 the general secretary called to speed up the development on unmanned intelligent combat capabilities. the efforts from beijing were bolstered by its military fusion program which allows the pla to harness commercial sector technologies for military uses. with regards to implementation on the pla, beijing is engaged in military ai research development and experimentation but so far open source

information about the pla fielding the military systems of scale remains very sparse. china has a large and sophisticated drone industry for the civilian and military sectors. but those systems appear to possess only partial forms of autonomy and cannot yet execute the most advanced types of autonomy that would be enabled by our artificial intelligence.in other words, they still rely heavily on human operators. as the department of defense 2023 chinese military power report says, the pla is pursuing greater autonomy across a range of systems for various battlefield purposes. regarding ai and the battlefield functions, china is already using ai for cyber applications where the technology is mature as well as intelligence surveillance and reconnaissance or isr tasks that have high levels of automation. the next stage for a implementation in the pla in this part of the pla will

likely be adopting the technology for low risk uses like logistics, maintenance, and training and particularly when similar commercial ai systems exist and can easily be adopted for those purposes. eventually it could implement into more of its decision making decision-making and command and control functions to move towards what chinese analysts call a command brain where machines and humans are seamlessly integrated. those application would nest necessarily start at the tactical level and only move up to more complex task at the operational strategic levels later run. it is important to note that china could fall short of its military ai ambitions for multiple reasons including technological shortfalls, a lack of trained personnel, bureaucratic competition, rampant corruption, the ccp need for tight control and even

a lack of funding as competing priorities compete for resources. to be clear, none of these obstacles mean that we can be complacent. far from it. they should remind us that effectively developing and fielding the cutting-edge of military technology takes more than brand plans, even for china. that said if beijing manages to overcome those obstacles, military operations could pose a risk to the united states in several categories. the first is potential shift in the military balance of power as small improvements level up and later on that china achieves the fundamental breakthroughs in military ai, given the advantage over the united states. the second relates to autonomous systems that might have more capabilities or which beijing might see as acceptable to use because they don't risk human operators directly. the third area of risk is in command control and communications. both through improved c3 and

conversely the use of ai to degrade u.s. and allied c3. the capabilities allow nuclear tracking or counterforce options or if states integrate ai into their nuclear complex in dangerous ways. those are the risks. i will conclude with recommendations for u.s. policymakers as they respond to the chinese pursuit for military artificial intelligence. first, -- do so in a relatively narrow way that avoids self- defeating steps. two, build military ai capabilities to stay on the cutting edge. a key part of which would be robust testing and evaluation to make sure military and systems are safe, reliable, and effective.

third, continue to shape mobile rules, norms, and institutions around the deployment and use of military ai. i would note here that china is already actively trying to shape the rules here and proposed a global ai governance initiative last october. fourth, the u.s. should engage with china and clear the way on military ai risk and then fifth and finally, the u.s. should prioritize intelligence gathering and analysis on an assessment of chinese military ai capabilities, which i said are evolving quite quickly. i will wrap it up there. you go for your attention and i look forward to your questions. >> thank you. >> thank you for the opportunity to testify about how the chinese military views the prospect of generative ai for manipulation and on the evolution of cyber enabled influence operation efforts.

i am honored, as a former commission staff. i want to make three key points. first, generative ai has the potential to revolutionize cyber enabled influence operations and supercharge maligned ability to undermine the democratic process in the united states and around the world. the key breakthrough is the dramatic improvement in authenticity and scale at a lower cost while also reducing human labor requirements and the probability of detection. second, the pla is already known to be one of several chinese communist party actors engaged in cyber enabled operations. and there are many reasons we should be concerned that the chinese military will incorporate generative ai for improved effectiveness and skill. third, congress and the broader government need to be prepared to live in a world with a much worse informational environment. more pervasive and more maligned influence operations.

i want to briefly touch on three main topics. the chinese military strategies, capabilities and intent, as well as the potential policy options for policymakers. first on the strategy. the pla has targeted and tailored influence operations and social media gave them targeted i/o and generative ai -- i open give them that. the overarching operational concept is cognitive domain which is described as combining psychological warfare with cyber up liberations to shape adversary behavior and decision- making. with the likely intention to deter u.s. and third-party countries into future conflict or as an offensive capability to shape perceptions or polarize society. cognitive demand operations is a topic logically driven update compared to the content you might be familiar with. it fundamentally reflects a shift as to how china thinks

about the traditional space of air, sea, and land based in savory. this view is not yet formally pla doctrine but is is important for congress in the community to watch. tentative ai fits in cognitive domain operations by supercharging the pla performance but so far there does not seem to be a shift from pla tactics. they are reviewed for this testimony. pla researchers recognize the potential for generative ai to improve content generation and contact distribution and recognize benefits for scale, reduce cost and more. i observed researchers using three. first, influencing public opinion the via bot networks. second producing intentionally biased publicly available

models. there, specifically degrading support for adversarial leadership this aligns with existing explicit interest by pla researchers to produce inauthentic contactor synthetic information as well as run large-scale --. i want to know that pla writings on the topic of generative ai and influence operations express concern about the potential use by the united states against the ccp in so many words to understand the public reports on the u.s. government activities. especially those of attributed to the department of defense are closely watched by pla members. this is not just academic. some pla researchers call for adopting cyber enabled i/o to respond on the topic. so far we have no direct evidence that the pla is specifically adopting ai for its enabled operations however, i argue that the pla is currently capable of doing so

if it chooses. moreover, a recent report highlights the pla affiliated researcher who has been working on a system for online public opinion struggle since at least 2016 and has been envisioning an into and automated online influence system since 2019. basically the technology he envisioned in 2019 have come through with ai breakthroughs over the years passed. other party state actors have been adopting ai and early evidence suggests it may be improving their performance. lastly, on intentions, generative ai is his supercharged objectives such as shaping foreign and public and domestic public opinion, deterring u.s. involvement in the conflict and degrading -- among other objectives., however, were to specifically flag the risk. a recently declassified intelligence report on foreign

election interference in 2022 found that china tacitly approved efforts to try to influence a handful of midterm races involving members of both u.s. political parties. and the odi worldwide threat assessment said that beijing, quote, has shown a willingness -- that involve perceived anti- china politicians. " a 2021 article by pla researchers suggest at least some in the pla are already using social media to identify politicians who are pro-or anti- china, in my analysis, to back election interference. to use twitter data from sitting u.s. government officials with no -- with known views of china to train a model and then use it to predict how other u.s. politicians view china and validated the results by consultant does consulting u.s. intelligence analyst. in their own words, the capability is intended to

assist intelligence analysts in their assessment of future past u.s. and china relations. i will highlight several policy implications. first, require social media platforms to label generative ai content and redouble their efforts to cut to combat they could counts. second, commit now to publicly releasing a nonpartisan declassified assessment of the u.s. intelligence community following the u.s. 2024 election. , encourage taiwan to increase information sharing publicly and privately about chinese cyber enabled operations and support the taiwan engagement with other democracies to share lessons learned and best practices. to explore the possibility for agreement this, conduct an end assessment thank

you again for the opportunity to testify. i look forward to your questions. >> mr. parker. quantum technology is an emerging technology by the highest levels of leadership within both the united states and the prc government. most quantum technology applications are still early stage and there are still many unknowns regarding which ones will eventually prove useful. but in the long run, quantum technology could greatly improve the abilities to collect, process, and transmit information with significant applications for national security and economic prosperity. specific potential applications relevant to national security include positioning and timing without gps. materials science and decryption.

quantum science is a highly -- the focusing on the industrial base of the nation's does not capture the complete picture. the united states and china are the leading nations in quantum technology by most metrics including patenting, scientific publishing a demonstrated prototype systems. to summarize a complex story i would say that of the three main quantum technology areas, the united states leads the world in two. quantum computing and quantum sensing while china leads the world in one. often communications. the two nation governments appeared to be pursuing different r and d priorities. i would say for the most important national security of boston tec identified as a low priority. overall, i'd say that for the

most important national security applications of quantum technology, chinese scientists are impressively fast followers, but rarely at the true forefront of innovation. the united states has the good fortune to lead it in almost all of the most important areas of quantum technology. moreover it has a huge advantage that the prc does not. a network of close alliances with many of the other leading nations in the field. neither country's military appears to have integrated any technology other than atomic clocks into actual operational systems. but the u.s. military has begun field testing quantum sensors and clocks in operational environments. that having been said, chinese researchers appear to be fairly close behind in a few important areas, such as super conducting quantum computing. the highest impact quantum technology applications are probably still at least a decade away. so u.s. policymakers should not get complacent because there is still plenty of time for global

technology leadership to change hands. at the csame time, i caution against framing quantum technology development in entirely zero sum terms. the u.s. and china are the strongest research collaborators in quantum science. valuable information and talent will flow in both directions. the biden administration recently issued an executive order restricting u.s. outbound investments in chinese quantum technology firms. i am not aware of any u.s. firms currently investing in chinese quantum technologies or considering doing so. moreover private industries will play a relatively small role in the chinese quantum ecosystem. most twill occur the national laboratories. so i suspect that the executive order will have little impact on chinese quantum rnd efforts on the near term. i'd like to conclude with a few thoughts regarding steps the

u.s. government could take to strengthen quantum technology. if they determine the nt technology should be a strategic priority, then the most important step towards that goal would be to continue to invest in fundamental science research. another important step to strengthen the work force including the domestic and foreign pipelines of skilled talent, while ensuring appropriate protections against intellectual property loss are in place. also broad export controls on quantum technology would run the risk of slowing progress and stifling a commercial industry. i believe that targeted export controls on specific chinese organizations of concern are low risk. but congress should carefully consider the impacts on the u.s. commercial industry of any proposed broad export controls on quantum technology unless those expert controls are directly tied to a concrete

military capability. finally there are three aspects of the ecosystem, that the u.s. government should consider understanding and monitoring. the financial health of quantum technology companies. the flows of intellectual property between the u.s. and competitor nations. and the supply chain for critical components and materials. i do not see any clear needs for immediate actions regarding these topics, but all three represent the potential risks to the long-term stability of the emerging quantum ecosystem. thank you. i look forward to taking your questions. >> thank you each of you and we will begin with questions, going in reverse alphabetical order, so i get to go first. jacob, let me ask some questions and appreciate your testimony. the first is and maybe

something that you have not looked at. i've seen recent reporting about ukraine's use of a.i. on battlefield in terms of target acquisition, etc. number one, have you looked at that at all or are you aware of it? and two, aware of any chinese efforts to study what ukraine has done and how it is being applied on the battlefield, so we have a better understanding? >> absolutely. the war on ukraine has served as a laboratory for a lot of these emerging technologies. it is certainly including china. not just how ukraine is doing it, but how russia is responding and the technological kind of thing they have going on the battlefield there. so it is certainly an area of key interest tifor china becaus they have not fought a conflict in a long time and certainly not with these emerging cutting- edge technologies.

one particular area that i'm concerned about is that russia in the content is gathering a lot of data from their systems in the course of fighting that conflict that potentially could be useful in the china, russia relationship, that the kind of data or insights on technology drawn from that might later be transferred to china. so i think that is something we will have to watch closely as we go forward. if you break it down, there's four parts of it. there's algorithms, data, and the talent to make all those things come together. so if we think about their progress, we can look at each of those component parts. certainly in the data and probably in the algorithm space, this is a place where they might be able to gain through some of the advantage indirectly through russia. >> and also how would you

assess u.s., and i know some of this is in your testimony, utilization, for example, predictive maintenance and some other opportunities from a.i.? how would you judge u.s. capabilities and implementation and, again, you also talked about it is still early for china. but what visibility or lessons might we gain? >> absolutely. i think the e u.s. has been ver focused on this in a pretty bipartisan way with things like the development of the artificial intelligence office in the department of defense. recently announced their replicator initiative focused on crude systems in particular and also their strategy on data analytics in a.i. adoptions. so there is certainly a recognition of what's at play here. i think that the challenge that it will come up against is the challenge that many other areas

of weapons procurement in other systems procurement will meet, which is often our procurement system was built in the 20th century. so it is not always well positioned to leverage technologies in the 21st century. i think there is a lot of energy in dod trying to get ahead of that, but the obstacles are pretty big. and so that is going to be a place where we move from experimentation to application at scale, we are really going to have to push forward to keep up with china. but also thinking about how to do that in a way where the systems are safe, secure, reliable. what is testing an evaluation look like in particular in systems that can change over time as they learn. so you can basically assess the system when it is built. but t then how do you monitor i over times? there is some unique challenges and testing evaluations for systems that we are still working to get our arms around out in the services as well.

>> thank you. >> and let me -- connect the last panel on this one and get your thoughts, you know, the last panel a enlot of it was about data collection. it was one component that was data collection. for generative ai and influence cognitive impact, etc., there's the broad application. there is also the targeted application. the ability to target certain cohorts, whatever you will is significant. how do you view able to gain large edata sets and everything from gee owe location to particularized organization? how does that fit about your

comments with generative ais and potentially other disruptive approaches? >> thank you. i think that's a great question. i would certainly say that fundamentally they recognize the importance of data. they will talk a lot about data from the open source perspective. it is really lydifficult to loo behind the curtain to see what is the quality of their data? how long do they actually leverage their data? it is an interesting question and it is hard to look at in- depth from the open source perspective. but to your specific question on, you know, non-u.s. owned social media platforms, that is absolutely something sthat cou come in to play for this and in looking at chinese military writings, i have not seen much specific discussion of tiktok or, you know, the gee owe location data, etc. but the theoretical risk is certainly there and that is

something to consider. and i also note that beyond tiktok, organizations have been collecting vast throws of data from not just u.s. citizens, but globally. there is a question of how much the researchers or the actors would have direct access to that data, certainly again. it is difficult to really make that assessment. s but theoretically, that will actually support their efforts to improve their use of generative ai. >> and i will just say that i think we have seen over the years, many years, some of us on the commission that theoretical risk will turn into real risk in a short period of time. and so if we can think of it, they have thought of it and they are probably putting it into practice. commissioner shriver? >> yes, it is a very key point and that you have described it

as potentially shifting the military balance in their favor. but some of this will be difficult to track from concept and rnd and development of particular capabilities to understand the progress that they are making in terms of application and seeing it in the training for the particular contingencies or nothing, you know, is this a disinformation point that is a result of generative ai or sort of classical, you know, what we always see. i had a couple specific questions on recommendations and then a follow up if i have time. mr. stokes, your recommendation number one was bold action to this wart thwart the applications. i i don't think that i noted an specificity under that. are there familiar things that you would have in mind that would constitute bold action in this case? >> yeah. i think that here, we would

have to think about, well thank you for your question. i think we need to think about those categories that i described earlier. if you think back to october of 2022, semi conductor, broadly considered those are fairly bold and they were updated in 2023. but there is still someone described them as leaky and that is a good way to think about them because there is still a black and gray market around them and to access that kind of computing power. and also china is making domestic advances in the manufacturing capabilities. so that is kind of one node. we also need to think about how we handle things like open source algorithms, where some of those advances might be accessible to china, the data, and then also the talent question. so i think that we will have to, i support the chip controls because they were the right level of boldness and kind of in many ways, cutting the feed

out from under china's domestic semi conductor manufacturing capability at the high end. they were limited in part because they needed to bring allies and partners on board, specifically japan in the netherlands, where a lot of the most advanced chip making equipment were made and they need to strike the balance there as china responds and reiterates on how they respond to those kinds of controls. i think it is really about opening it up. so for example on cloud computing, my understanding is the commerce department has put forward, i think it is an interim rule that is requiring u.s. cloud computing companies to report who their customers are. thinking about that, probably that against the list of prc military entities that were updated just yesterday. those kinds of actions were going to have to think about across the full spectrum of the ingredients for ai. >> thank you.

>> and sorry. your recommendation about social medias should be required to identify when content is a result of generative ai. is that simple to do? i mean no? is it simple to identify? i'm not asking, you know. is it easy to get such a law gr passed or something? >> i think that is a good question and there are two parts to this. one is social media regulations for congress, the u.s. government. that is mecertainly an ongoing conversation and challenging and somewhat fraught. but also as you mentioned, technological challenge of being able to identify the content produced by generative ai models. right now, it does seem that the offense in a sense has the advantage, right? that it is easier to produce this than to identify it.

you mcan get that identification, but overall, especially at scale, right now,. it really seems like offense has the advantage. a part of my recommendations, you know, policy options to consider was indeed for the u.s. government to support investing in long-term abilities to detect generative ai content. obviously it would be beneficial to find the public- private partnerships. they are very well placed to support the development of that technology, one because they have the data and really the first observer on these trends. and so there would be an opportunity for them to potentially work collaboratively to help do that. >> thank you. it seems to gme that it could b a low hanging fruit if we could do it from a technological standpoint. i'll ask another question if there is a second round. >> commissioner isvice chair price?

>> thank you. and thank you all for your important testimony today. i wanted to jump in where commissioner shriver left off. so when we are talking about understanding this and what the long-term need would be. what are we talking about? as we give your recommendations to congress. what are numbers? and then my second question following up on what we just said was how do we do this kind of media literacy training for the citizens or for others in government? h it is worth it on this front. second on tsupporting media literacy, other colleagues at random have done a lot of work

on the broader crisis of one of the frequent recommendations from the colleagues is investing in civic engagement, media literacy. so for example, dod employees who have a security clearance will need to go through an annual security data and do specific training. so far to the best of my knowledge, that training doesn't include learning the basics of identifying generative ai content. f and otherwise support broader u.s. government. and this is really a whole society challenge. generative ai doesn't make a new problem on media literacy, but really emphasizes the risks of the fshort comings so far. it is another opportunity for congress to support the outreach and really, you know,

improve broader u.s. kind of citizen awareness of these risks. and i want to give you a minute to expand on it a little bit more and why particularly? >> thank you for the question, commissioner. which i would describe it as areas where the u.s. government has limited visibility are first financial health of quantum technology companies. second, flows of skills and the talent. and third, the supply chain. briefly, i'll touch on all three of them. i think to the financial health of quantum technology companies, the technology is ti still very early stage. there are big questions as to

realistically y timelines for actual revenue generating applications. most industry investment if quantum technologies and computing, which is arguably the most technologically field within quantum technology. and there is a lot of venture capital investments and the activities. but they reported fairly modest of the revenue. and so there is a question as to the long-term financial stability of i guess the t technology readiness timelines are aligned with the current state of industry. what steps might be taken if there is, for example, a recession, which made the investment environment less favorable for so-called deep technologies. and secondly regarding the flow of skilled talent and the intellectual property. there is a lot of foreign

talent in the u.s. research ecosystem. many graduate students are from foreign countries. i think on the whole, that's a good thing for the u.s. research enterprise, but there are questions as to what are the appropriate protections for the property, not only amongst students, but also companies, which may not have particular expertise and cybersecurity, for example, small start-up companies. third, the supply chain that the technology is still very much there and there are different approaches being pursued, which will have different supply chains. i think there is a general lack of understanding as to where some of the most important components are coming from, which countries they are coming from, especially as we go several silevels down into the supply chain, whether they are coming from the competitor countries. how robust that supply is, and whether they are diversified across multiple vendors.

for and what a mature quantum technology supply chain will look like. all these of those areas are areas of we are just more information that would be helpful for having a perspective on the ecosystem. even if immediate actions don't necessarily need to be taken. thank you. >> thank you. >> cochair hilberg? >> thank you, co-chair. i kindly ask our witnesses to keep their remarks as brief as possible in the interest of time does the pla cognitive warfare land, air, space, sea? >> yes. >> how much do you believe the pla has invested in cognitive warfare? are there public estimates available? >> no public estimates available. >> do any of the liother witnesses know of the public influences? and are you aware of, do you believe the pla is actively working to train models to

enhance the way they target americans through the information operations and its cognitive warfare efforts? >> that's a good question. i can't point to specific, you know, public available research directly discussing training generative ai models. right now the publicly available research that i surveyed from my testimony was more on i would describe kind of the theoretical side of understanding the value. but i would absolutely not be surprised if at least some of the pla were beginning to experiment and work with these models, yes. >> i think it would be prudent to assume they are, but i don't think we have good open source information on that topic right now. >> are either of you aware of bytedance rivals opening up lego, and doesn't bytedance own

tiktok? do either of you believe there is a real plausible risk that bytedance is using tiktok data to train their lego? >> i don't have in-depth knowledge on bytedance. but the possibility of any company will want to train their algorithms on the best data. >> yeah, they are working to build the types of those because they see them as a market opportunity for themselves. if they can access the data, i believe they are. >> yes. >> aren't they dual use in

military context as well? wouldn't it be common sense for this country to bar the chinese social media applications from harvesting, collecting vast amounts of american user data to train, advance ai models that could be repurposed to kill americans? >> i think it's a challenging topic and worth considering whether the foreign owned companies should be able to harvest the data of u.s. citizens. >> we're talking about social media companies under the control and influence of a foreign adversary government. your response applies to those types of companies, is that

right? >> yes. how high do you think they rank ai in their modernization efforts? would you say we're at a reflection point in the history of warfare and that they are poised to change in the years ahead? >> i believe they rank it pretty high. there is intelligence and they say they have reached mechanisms, so working on the next stage, but increasingly trying to integrate intelligence. so the last revolution of military affairs with the coming one sort to speak. whether they are able to do that is the question. but i think that is how they characterize and understand the environment. >> what do you see as being the most important thing for national security? how should we respond to their embrace of ai in this intelligent concept that you

described? >> we should think about it along three lines. one is improving our own capabilities and making sure we could keep pace. two, how do we slow down their capabilities, especially where it applies or the most dangerous applications thereof. and then third, working domestically and internationally including in some cases with china to shape the rules and the norms around controlling this technology and it has to be a three-part strategy. >> can you explain the first one about enhancing on capabilities? >> yeah. i think, you know, just the initiatives that i described earlier undergoing, underway at dod. you know, i think that our recognition that warfare is changing and that it has the potential to change with these new technologies and making sure that we are staying on the cutting edge. not just of the basic technologies themselves, but also integrating them into our institutions, having the right

personnel to train and operate them. and really across all the dod enterprise. so i think that really how we should understand what we need to do in that pillar. >> and do you see the adoption of military ai as focused on the particular part of the kill chain, like logistics, targeting, or is it focused on all parts? >> i would say it is probably focused on all parts, but there might be more emphasis on trying to, you know, to have the approaches, so we can, you know, not just have to kill the system sort to speak, but to be able to disable their operation. increasingly as we would go towards a more integrated commanding control architecture like for the u.s. china is thinking in those same terms. their term is multi-domain precision iwarfare verses the

precision warfare. so it is really across the kill chain i would say. >> my last question is for mr. parker. as commissioner mentioned, we are still in learning mode when it comes to technology. can it be used to conduct the underground mapping? sub surface is an area that the u.s. dod acknowledges and that they have acknowledged that things could be used, for example, the underground tunnel detection, yes, in principle. the technology is not necessarily there yet, but certainly theoretically possible. >> thank you. >> commissioner glass? >> many thanks to you all. my first question is for mr. stokes. to what extent do you think we are gathering intelligence with

some of our international allies around chinese use of ai in the military space? is it adequate enough? if not what kind of risk association should we do and how much prioritization should we give this? what can congress do in that regard? and so that is my first question. and to nathan, sorry, i'll mispronounce your last name and my apologies. you know, you mentioned about ai and the social media influence, the influence campaigns, especially as we have an upcoming election this year that what could congress immediately do to help safeguard some of the concerns that you raised in your testimony? >> thank k you for your questio commissioner. i think that we will have to draw from our closest allies, but really across the world. even is probably going to have

a little bit of a different, you know, they have a different hand on the element to use the metaphor on what's going on with china in this particular space. and i think that broadly speaking, we do need to make it a priority given its importance in terms of our intelligence collection. i'm not an expert, but broadly speaking, you would see in the u.s. intelligence leaders, talk about the need to change the u.s. intelligence enterprise to catch up with some of these emerging technologies, and that the way we would gather the intelligence and analyze it that they might have to change. and i can't imagine a more important topic for the security of the united states to make sure that we would get our understanding of where china is especially in military ai that we would get that right. >> i think it is a key question this year. i'll give you four pillars to

think about. first as i was talking about with the other commissioners before, considering social media regulations or at least engagement that it could be positive and collaborative to make sure that social media companies are well positioned to do what they can on their platforms to identify. and you know, if appropriate, remove inauthentic content. second is to support and really empower as much as possible the u.s. government transparency on these topics. they could only provide so much information and to have so much information and sometimes u.s. government will pohave addition information as seeing the declassified assessment on 2022 elections. but transparency is really important for supporting u.s. public faith integrity. third is as mentioned before, civic engagement media literacy. that they will have a role to play in both supporting the

efforts, investing the efforts. in terms of the services. that is really something that congress could play on as a key touch point for the u.s. government with the broader american public. then lastly, i do think that it is worth at least having some engagement with beijing on the topics. i saw a cnn report i believe yesterday morning, that xi jinping has promised biden, i believe now twice that he will not interfere in this year's elections. fbi director wray testified to the select committee yesterday was asked about it and said basically i'll believe it when i see it. i'm old enough to remember when xi jinping promised biden. so i'm heartened by the promise. i think it is very good that the biden administration is at a high level engaging with beijing on the topic. and as an analyst, i'll look at capability and intense and e equals risk or threat, and i

think both are there. this kind of promise, is xi jinping going to follow through, question one. is the rest of the party/state going to fall through? he doesn't manage it day to day, right? third, how do we define the election interference? and that i think there are a lot of challenges to that and good to see the biden administration engaging. the last point i'll make as jake mentioned briefly, chinese government put out a white paper ahead of the u.k. summit in october of last year on ai global governance. one of those tenants said that no one should be doing the driven manipulation. i believe that is low lying fruit for them to say great, we agreed. let's not do it and so far i have not seen them seize on that as an opportunity.

again, we can discuss on whether we trust that and it san opportunity for positive engagement. >> thank you. >> commissioner? >> if i could continue with you. you said in your testimony written and you said it again. a this morning that they fear that the united states in the west would use generative ai to undermine it. and taking place. do you think this is an expression of the general concern? and that they might like to be able to do? >> and thank you. >> and it is the aspect of thinking. and so when i'm reading these pla writings, they do talk a lot about the foreign threat, sometimes they as western threat. sometimes they specifically

address, you know, u.s. intent. this is a a part of a long standing broodier ccp concern of u.s. information operations. and nearly every single chinese domestic crisis at some point in time has been blamed on the united states. hong kong protest, we can go on and on. and the ministry of foreign affairs has kindly put out several white papers over the last several years that neatly and very long form lists their view of u.s. government activity. and so it is important to understand that it does not mean we have to agree with it, but understanding the perspective as an analyst that it is important. and what really drew my concern is that some of these writings described it as wa tip for tat and in response. the report to congress last year said the same thing, right? and that i quoted it in my written testimony from a

perspective, other countries who are engaging in the influence operations, basically justified a response. so i think that it is really important to understand. >> thank you. >> and you referred in your testimony to the supply chain for quantum computing or devices. could you say more in what is involved that it will be made out of the conductors that will go in to all the other computers and who will make these things? >> thank you for that question. it is correct to say that it is very, very different from traditional computing technology. that they do require a lot of computers in them for control.

and they often revolve around the low temperatures. so surprising amount of quantum supply chains that will revolve around them to get down to the degree of absolute zero in some cases. and these technologies are unusual and very critiquing. some involve around lasers. others revolve around refrigerators. most come from allied nations i would say. relatively little comes from russia or china and they do not have that capacity either. and it is one example in dilution e refrigerators that it is controlled by the single company. on the laser side, certain companies in japan and germany make very high-powered, high- quality lasers of the frequent sis, which are difficult to

source in other sources. so it is a difficult thing to categorize globally and to say that there are many technologies that are sourced all over the world. and because there are supply chains from the difficult technology approaches that many may not be important and that it might be one of those approaches and that they will win and that will become extremely important strategically, where the other six or seven technical approaches may become irrelevant. so we need to sort to track all of these supply chains. >> thank you very much. mr. stokes, you said it was important if there were going to be limitations and controls imposed on artificial intelligence that they would

focus on military applications. and with the technology that is often describinged as being equivalent to electricity in their general application in the early stages of development. >> i think that it will be more about targeting building control to control what we can. and some of the domestic industry. and so i agree with you that it is very hard if not impossible to control military ais or commercial sector ae application for military especially in that context. it is more about trying to understand the dynamics. and that the advantages are in that account for that and so i think that sometimes it is just to shut everything off.

and that there are good reasons to feel that way and effective over time. to consider the particulars of the given market that they were operating in including what will drop the american companies. >> thank you. >> acting chair cleveland? >> please emphasize that. the acting part. >> i have a similar question to dr. freedberg's that i'm interested in the supply chain. you say in your testimony that you would talk about refrigeration in lasers. you just said that there are six or seven technologies that might be relevant. and you don't know which ones the likely winner. so investment in all right now makes sense.

are any of those technologies controlled by any export regime or arms control? are they subject to any kind of restriction in terms of transfer to china? >> there are no broad expert controls that will apply to quantum technology specifically. now there are a couple of caveats there. one is certain specific chinese organizations that have been placed on the commerce department's entity list. so several of the leading chinese research organizations do not have legal access to u.s. exports in those. the second caveat is that there are some expert controls that apply to certain types of sensors broadly, which are not specifically taylored to the quantum sensors, but would apply to quantum sensors. just for example the control. some might apply to magnum. i think if the united states

decides to impose expert controls on the more mature technology, it would be useful to update those controls, which you know, may or may not apply, it's a bit of a gray area because they were crafted before the current technologies became more mature. so it could be an area that will use a little updating in clarifying. >> i appreciate that. and i think that you also said in your testimony that much of this technology is made in japan and europe. and so it would require collaboration if it is to be effective, i would assume. >> yes. i wonder if you could supply for the record since none of us really understand quantum. and kind of your assessment of what the technology needs to be and just so we would have a very primary understanding of what we're talking about here and that i would appreciate that. you also said in your testimony

that alibaba, ten cent, that they have invested in quantum, and i assume that there is some leader that it was designated to undertake this effort as the government usually designates a leader. but you say that they appeared to have pulled back from that field and they shut down the quantum computing research labs in november. talk a little bit more about what may be behind that, if not, trend. series of actions? >> and so that's correct. the chinese quantum ecosystem will look very different from the u.s. one. the u.s. ecosystem is by in large led by the private industry at this point. the chinese quantum industry is much more heavily funded, focused in government funded national labs. and already the sort of weight is significantly heavier in national labs in china than in commercial industry. several of the large chinese players that you mentioned by

tencent, etc., had quantum computing efforts. and they have several of them, if not all of them, have shut down in the past six months by doing that. they were selling all their quantum hardware, quantum computing hardware to national lab. that concentration to the national lab thwill appear to b consolidating even more in just the past six months. i don't have great visibility as to why, but i do make that decision. and my guess would be that tay would assess that they were not particularly competitive in that space and they were far behind the u.s. companies that did not seem to be catching up. and they did not see it as a revenue generator. so the extent that we could talk about chinese thinking as m a whole would be further doubling down on the national laboratories. none of the chinese quantum technology companies seem competitive globally ever. >> so you would characterize it as an increased risk that they are consolidating under the

government control and these capabilities teor along with an assessment by these companies that it is not profitable? >> i don't know if i would necessarily characterize it as an increased risk. i don't think y it really matters too much from the u.s. national security perspective whether it is the chinese government and the industry that's controlling the technology. i think in china, they are more or less the same thing. i would maybe interpret it as cautious good news that the chinese at least one aspect of the chinese ecosystem does not appear to be self-assessed, but it is not at the very top of the global competitive landscape there. i don't think that it is a particularly major increase or decrease in risk either way. >> thank you. >> are we going to go to a second round? >> we will try to keep this round shorter, so that everyone has a chance. i'm going to pull a star trek

issue here because it has baffled me. but while we were talking, i looked at it again that china is rumored to have successfully engaged in threaperration. teleportation, which was seven years ago. you're nodding your head yes? i assume that you're aware of this? i believe science magazine and others would talk about the profound nature of that. and that is seven years ago. number one, has that been peer reviewed and is it true? number two, what are the implications? i can't wrap my head around that. >> thank you for the question. it is a difficult question to answer briefly. but i will try. [ laughter ] and the quantum

threaperration -- teleportation, maybe iit was in europe. it could be useful for networking together quantum computers. in the near term, there are no immediate concrete applications that will improve transformative. but i think tit speaks to a larger part usof the chinese focus and prioritization in quantum, which is they're focused on communications technology and the use of quantum technology to secure their communications. some disagreement on how significant that is operationally, whether it is actually useful capability. but the chinese do appear to have chosen that to be an area in which they want to excel globally. one aspect of that is china has launched two different satellites that are capable of quantum communications from outer space, mediated by satellite communication. no other country has known to

have launched the satellites. again, the u.s. national security agency actually publicly said that specific application is not a high priority. not something that had they intend to pursue themselves. so there is definitely disagreement on how seriously we should take this and it is one area that they will be leading globally. and just one last technical note. the teleportation does not allow teleportations of humans or significant quantities or materials. >> i read about it in terms of quantum internet and how it might enhance command, control, survivability, etc., in terms of military facilitations. so not expecting that he's beaming up somewhere and let me say that for the record too. commissioner shriver? >> dr. parker, i think you forgot to end your sentence with yet.

thank you. both mr. stokes, he mentioned the need for dialogue. >> and on these matters. i think nathan, you already had a chance to expand on that a little bit. se so mr. stokes, i'm wondering how you see that as a useful engagement in how you would, what topics you would prioritize, etc., noting that this is an inner locking and the weather balloon that went off course and that they know they shot it down and collected it. and what are our expectations be in such a dialogue? >> sure. thank you for the question, commissioner. and i think that it will psbe important to note about the timing of this and that it on comes after major steps in the u.s. to put it in place of the regulatory framework for our own ai and also to work internationally through the moves like the political declaration on the responsible

use of military artificial intelligence and autonomy. and trying to set a framework for what we think normatively the rules ought to look like in ai and then engage china from that perspective. i think it is worth, you know, stepping back and saying one of our r grand strategic objective is to support our rules base international order. the rules are being written in real time here. and so we need to think from that context. and it will make sense to build china in if we could get them to accommodate and agree to a certain set of rules that we'll put in place, leading the coalition in defense of. as it relates to the specific as talks, the national security adviser, sullivan, he talked n earlier this week that they will happen some time this spring? it will likely be very basic about exchanging views on what

we would see as the actual risk from ai technology, and understanding more about how china understands the risks. and maybe in certain areas, sharing information about how we would test and evaluate our weapons and what parts of particular the nuclear command and control complex would or would not have artificial intelligence. those types of very basic topics are where you would have to start here. you would be miles, perhaps even decades away from an agreement that would look like an arms control agreement where there is verification and those kinds of things. both for the political reasons that you would raise, but also underlining the technical challenges around the verification with ai. so i think it is a good thing in that right context and with the right low basic expectations. thank you. >> vice chair price? >> thank you. i just have one quick question for mr. stokes.

on your recommendations for policymakers. your last one you say we should prioritize intelligence gathering and analysis. are you cujust flagging somethi or are you suggesting that congress ask for something in particular? >> i think it would be beneficial especially for the open source analytical community that i work in to have more open source information about this. you know, one potential way of doing that, that congress could do is to mandate a certain, an additional section in the china military power related to some of the topics as they have done with the other topics in the past. you could also probably mandate that the dod or the ti intelligence community will provide briefings on this particular topic to members of congress in relevant committees and even commissions. those kinds of actions, because i think we are at a place in the policymaking process where we are still getting up to

speed on what the technologies are, what they do. and then trying to get past that to what should we do about them? and having an accurate understanding to the extent that it is possible of where china is relative to the u.s. in a robust way would be really helpful. so that is where i would start as a recommendation within the chinese military power report. >> thank you for clarifying. that's it. >> cochair? >> no additional questions on my end. >> commissioner glass? >> you mentioned translation, i think, of the phrase, something like controlling the brain. chinese have several categories and things that they talk about. is that simply a figure of speech, you know, to influence your thinking or is there more to that? >> thank you for asking the question. happy to clarify.

so that was a reference to one of the popular kind of overarching conception of some of the things is that they will underline and often associated with cognitive demand operations, right? and that it is the most popular operational concept per se. when the chinese military researchers will discuss influence operations more broadly. so there is some discussion about controlling the brain. and that really has a long standing history in chinese military that they will influence the operations, psychological warfare and the operations, discussions. some of the conversations that you could find in chinese military writings are, i will say, futuristic. that is certainly not unique to the chinese military when you look globally about what some in global military's right about future potential breakthroughs and brain science

and other things. but one of the more concrete examples that i provided is actually something that chinese military researchers would look to the u.s. for, which is laser weapons and also acoustic weapons. so the ability to use that to some of these cognitive functions and to make it comfortable and ideally they would drop and lay down their arms, surrender on the battlefield. perhaps they would develop something the area of denial weapon. i said it in my testimony. so there is some conversations in writings that are futuristic and some that are more concrete and not totally, you know, distance from what u.s. dod has considered sometimes. >> okay, thank you. >> futuristic, and we just talked about teleportation. okay. acting chairman? >> yeah, yeah.

so -- mr. stokes, for decades, a part of u.s. military policy calculations have been the chinese military doctrine is very hierarchy and that a part of our strategy has always calculated or presumed that -- that decisions will be made by the top leadership in beijing. and i think, if anything, over the last decade that idea that xi is in charge and in control and is the ultimate decision maker on everything has -- this commission certainly has reinforced that idea. i'm wondering, mr. stokes you say in your written testimony,

there is some military leaders who might not trust applications of ai because it loosens that, that control. i'm wondering how ai is changing the idea that decisions may not be made in real time by beijing. does that question make sense? is there a shift in thinking that more control will be yielded to the field commanders or how is ai changing military, our military thinking about their military doctrine? >> well thank you acting chair cleveland for that question. i think -- my sense is it can kind of work in one of two ways. on the one hand as i wrote in my testimony, you can see a reluctance especially from mid- level commanders who will be held accountable for what these systems do and to employ them.

and sometimes they fail in the unexpected ways or sometimes you don't know what it is going to do. when you're responsible for both operational affects, but also the political implications of that. in other words if you're in the pla and you've got your politicals sitting right next to you, do you want to roll the dice on whether the system is politically correct in addition to being operationally correct? that's one thing that might hamper deployment of pla or military systems in the pla. i think on the other hand you see a lack of trust from the top leadership all the way down that they could implement decisions well. i'm sure you are all familiar with the concept of the five incapabilities. i think if you had too much faith in military ai systems or ai generally that the machine will make the right system and the machine is reliable. you might be able to have the notion that you could make all

the decisions from the top and let the machine push it down through the system. i don't know where that -- where that nets out. i think we're probably still too early to tell. thank you. >> could i add to that? i agree with jake. i think it is very early to tell. i think it's absolutely something very important to watch from what we can observe open source, not just in the writings, but as they're doing their exercises and training. it's absolutely something as jake mentioned, vital to engage with, china in specifically and china military on how they assess trust, and the quality of their ai systems. is there an opportunity for at least making sure they don't trust the wrong system in the wrong way? i think as jake said and as this commission has heard before, right? xi jinping has dedicated it. and so one of my concerns is, indeed, xi jinping decides

after some amount of time these guys and they are mostly all guys, can't do their job. then someone else offers him a magical ai system and he says great, can't be worse than these guys. i think it is absolutely a concern. i think it is something toll watch. again, there is hopefully some value and dialogue on this topic. we can't solve the problem for them, but it's something that's really important. >> what would we be looking for? if you were to say -- define some metrics for us to say if we see a consolidation what would we be kalooking for? can you define some metrics for us? if we see a consolidation of military leadership or the elimination of a layer of leadership, i mean how might we think about what would be a measure of xi xinping saying a.i. has to be better than what i've got.

>> nathan and i have had sessions together, we have a wall of metrics we can use to understand analytical questions like these. i think if we understand more about what systems we know are out there, being applied in the pla, we can probably infer from there about what level the control and orders are coming from, and how they are executed from there. we are still in an early stage, where we don't really know the specific systems, so i think i would start there, and look at things like training materials, and the way grades and specialties are named. those subtle indicators might tell us something about the way the pla your accuracy and command structure is working.>>

i have one quick question for mr. parker, and a quick question for the witnesses at large. you mentioned earlier that a quantum sensor could be capable of subterranean mapping. if you hypothetically attached a quantum sensor to a giant helium balloon that you guided over nuclear silos, wouldn't that reveal sensitive information about our nuclear silos?>> i don't know the answer that question in enough technical detail to comment. >> if you look at the overall pattern of behavior that you described today, and if you add facts about china that we know, the military drills, cognitive warfare, the adoption of a.i., wanting to reunify with tehran,

is china preparing for war? or is this the behavior of a peace seeking country? mr. stokes, maybe we will start with you, and work our way down.>> i think china is preparing to have the capability to fight a war. the intelligence assessment is that xi xinping has ordered the capability to invade taiwan by 2027. an order doesn't mean it will happen, it is a dynamic military balance, but it is clear that xi xinping wants china to be in a place to do that , to have that as an option. i don't think that necessarily means he will do it. i think he can still be deterred, with the right set of actions. it is incumbent on us, and allies and partners to determine what those actions need to be, and to work to

prevent any such war from taking place.>> i agree with jake's comment. i think we are really at the beginning of understanding of this idea, and there is still a lot of research to answer that question. it is a really important topic to understand, how does the chinese military see crisis, versus wartime? i think fundamentally, in terms of chinese leadership, i defer to the u.s. government on the assessments. i think deterrence is still a viable option. it's important to support efforts for deterrence, so we don't get to that point.>> i

cannot comment on whether or not china is preparing for war, but i would see the quantum efforts as a way to prepare for war in the long-term. they see certain areas of quantum technology as a big economic opportunity, a lot of it might be oriented around growing the economy. a slightly more specific comments, a lot of efforts in quantum communications refer to the goal of shoring up medications against external access. i think the provocation of a lot of that was the alleged leaks from snowden a couple decades ago. they probably said those alleged leaks reoriented thinking about the wisdom of relying on western communication systems for internal medications. i think that may have been a

motivation to revamp the communications system, which is not directly related to any near-term plan for warfare.>> a few all seem to believe that there might be plans for the use of military force in the not distant future, thank you for your responses.>> now to the commissioner for a quick follow-up.>> is there discussion in the open source chinese literature and military writings on the risks and dangers of autonomous weapons? >> i think both in the literature, and having spoken to jenny's experts on these issues, there is an understanding that there are risks involved, kata scott catastrophic risks for a.i. technology, and the specific wrist risk that a vehicle could

take an action that could precipitate a crisis that was unintended. i think there is a recognition, and that relates to the control question we were talking about. my sense is that there are times that probably wants to start a crisis, but at the time and place of their choosing, they want a system that might start an unintended crisis. i think they are aware of that risk, and probably see it as in their interest to avoid those kinds of outcomes.>> thank you to the panelists. we will take a break until 1:50 pm. we look forward to being back at that time. thank you.

[inaudible conversations]

[inaudible conversations] >> if you ever miss any c-span coverage, you can find anytime, online, at c-span.org. markers can guide you to interesting and newsworthy highlights. points of interest markers will appear on the right-hand side of your screen, when you hit play on select videos. this will make it easy to understand what was debated and decided in washington. scroll through and spend a few minutes on c-span points of

interest. today, a white house director and treasury secretary will be testifying on the 2025 budget request, and the economic outlook. watch the subcommittee hearing that 2:30 pm eastern. c-span is now on our free mobile video app. >> friday night watch our 2024 campaign coverage, a one-stop shop to discover what the candidates are saying, along with firsthand accounts from political reporters, updated poll numbers, fundraising data, and campaign ads. follow the campaign trail on friday nights, at 7:30 pm eastern, or download as