AI in America

AI as the New Arms Race

Competing with China

In the twenty-first century, artificial intelligence has replaced nuclear weaponry as the defining technology of global power. Once, geopolitical strength was measured in megatons and missile ranges; today, it is measured in algorithms, compute power, and data flows. The United States and China are the two principal superpowers in this new technological contest, racing not just for innovation but for ideological and strategic dominance over the digital future of humanity.

America's AI revolution began in the open halls of academia and industry; places like MIT, Stanford, Google, and OpenAI where innovation was fueled by venture capital, academic freedom, and a spirit of experimentation. China's ascent, by contrast, has been state-driven: a national strategy to achieve AI supremacy by 2030, blending government direction with corporate execution through giants like Baidu, Alibaba, Tencent, and Huawei.

For most of AI's history, from its founding as an academic discipline in the 1950s through the early 2010s, the technology remained primarily a research pursuit with limited strategic significance. Periods of progress alternated with "AI winters" when promised capabilities failed to materialize and funding contracted. Military and intelligence agencies maintained interest, but AI wasn't central to national power.

This changed dramatically in the 2010s as machine learning breakthroughs, particularly deep neural networks, demonstrated capabilities that seemed to promise transformative economic and military applications. The 2012 ImageNet competition, where a deep learning system dramatically outperformed traditional computer vision approaches, signaled a paradigm shift. Subsequent advances in natural language processing, game-playing AI, autonomous systems, and other accomplishments reinforced the belief that AI had transitioned from a niche research topic to a general-purpose technology with wide implications.

China recognized AI's strategic importance early. In 2017, China's State Council released its "New Generation Artificial Intelligence Development Plan," setting ambitious goals to match US AI capabilities by 2020, achieve major breakthroughs by 2025, and become the world's primary AI innovation center by 2030. The plan explicitly framed AI as central to economic development, national security, and global competitiveness. It committed substantial government resources and coordinated efforts across research institutions, technology companies, and industrial sectors.

The United States initially didn't respond. While American companies led AI development and US research institutions produced groundbreaking work, the federal government lacked a comprehensive AI strategy. Congressional hearings in 2018 and reports from advisory bodies warned that China's coordinated approach could erode American leadership. This prompted the Trump administration to launch the American AI Initiative in 2019 and the AI Action Plan in 2025.

AI's rise has created a new kind of arms race; not over territory or ideology, but over compute. The most advanced AI systems require enormous processing power, concentrated in data centers filled with GPUs and specialized chips. The U.S. leads in chip design, thanks to NVIDIA, AMD, and Intel, while China pushes domestic innovation through companies like SMIC and Biren to overcome export restrictions.

Washington focused on securing semiconductor supply chains and curbing China's access to cutting-edge AI chips. When the U.S. tightened export controls on GPUs, Beijing responded by accelerating its homegrown alternatives and investing billions into AI compute infrastructure. The result was an escalating cycle of technological containment and retaliation.

Data, the fuel of AI, is another front in this digital rivalry. The United States relies on open, distributed systems and consumer data ecosystems driven by advertising and e-commerce. China has access to vast centralized datasets under state oversight, enabling it to train models at scale and deploy surveillance-based AI systems across its population. Facial recognition, predictive policing, and social scoring illustrate the power of state-directed AI.

In the global arena, this divergence has shaped competing visions of technological governance. America's open AI aligns with democratic values compared to China's AI which is built for control and conformity. Nations from Africa to Latin America now face a strategic choice of whether to align with the U.S. model of open innovation or the Chinese model of techno-authoritarianism.

Both powers recognize that AI supremacy depends as much on human capital as on silicon. The U.S. still attracts top researchers from around the world, with many of them educated at American universities. Yet China's investment in education, graduating hundreds of thousands of engineers annually, and the repatriation of overseas talent have narrowed that gap. Universities, think tanks, and private labs have become the new battlefields of soft power, where research papers and patents are the weapons.

Despite the rhetoric of rivalry, AI's nature as a global technology blurs borders. American and Chinese scientists often cite each other's work, open-source code flows freely across continents, and companies like Microsoft and ByteDance quietly maintain joint research programs. The stakes rise as AI systems grow more powerful; capable of shaping economies, influencing populations, and driving autonomous weapons.

The "AI Cold War" is not inevitable, but it is already reshaping the international order. The United States seeks to maintain a lead by balancing innovation with regulation and openness with security. China seeks to leverage centralized control to accelerate deployment. The outcome will determine not just who dominates AI, but whose values are embedded in the algorithms and standards that govern the 21st century.

If nuclear deterrence was built on mutually assured destruction, AI deterrence may depend on mutually assured disruption. Autonomous drones, cyber defense systems, and AI-driven intelligence analysis have already transformed modern warfare. Whoever masters AI-driven decision-making will not only dominate the battlefield but also the global economy. America's challenge is to innovate responsibly by staying ahead without losing sight of the democratic principles that made it the birthplace of AI.

The AI arms race is not merely a contest of machines, for it is at bottom a contest of philosophies. One vision seeks to use AI to expand freedom and creativity, the other to perfect control and order. Between them lies the destiny of digital civilization. In this race, America's task is not just to lead, but to lead well by proving that innovation, transparency, and human-centered ethics can outpace authoritarian efficiency. The world is watching, and the algorithms are already learning.

In what follows, we take a deep dive into the AI arms race.

The Race to Global Supremacy

The competition between the United States and China for leadership in artificial intelligence represents one of the defining geopolitical contests of the twenty-first century. Unlike traditional military or economic competitions, the AI race encompasses technological innovation, industrial capacity, talent development, data resources, computational infrastructure, standards-setting, and the ability to shape global norms around AI development and deployment. The outcome will profoundly influence economic competitiveness, military capabilities, technological standards, governance models, and the global balance of power for decades to come.

This chapter assesses the current state of US-China AI competition in research and innovation, talent and human capital, data resources, computational infrastructure, commercial applications, military and intelligence uses, industrial policy and investment, regulatory approaches, international influence, and ethical frameworks. The United States maintains advantages in foundational research, leading AI companies, advanced semiconductors, and attracting global talent. China leads in certain AI applications, manufacturing integration, surveillance technologies, data collection at scale, and government coordination of AI development.

The competition is intensifying as both nations recognize AI's strategic importance. The United States has implemented export controls restricting China's access to advanced semiconductors and AI chips like NVIDIA Blackwell, while investing billions in domestic semiconductor manufacturing, data centers, and AI research. China has made AI central to its industrial policy, mobilizing government resources and private capital toward AI development, while seeking self-sufficiency in critical technologies. Both nations are competing for international influence through AI partnerships, standards-setting, and technology exports.

This rivalry carries significant risks. Technology decoupling could fragment the global AI ecosystem, which reduces efficiency and slows innovation. Military applications of AI increases the risks of miscalculation and conflict. The competition could accelerate development of powerful AI systems without adequate attention to safety. The different governance models, democratic versus authoritarian, create tensions over privacy, surveillance, and individual rights.

The competition also drives innovation and investment that might not otherwise occur. Both nations are mobilizing resources, developing talent, and pushing tech frontiers. The challenge is managing competition to avoid catastrophic outcomes while preserving benefits from innovation and maintaining possibilities for cooperation on shared challenges like AI safety and beneficial AI development.

Historical Context and Strategic Framing

The intensity of the US versus China AI competition reflects several factors that make AI strategically important:

Economic Competitiveness

AI is increasingly characterized as a general-purpose technology comparable to electricity or the internet. It is not a single product, but a capability enabling productivity improvements across industries. Nations leading in AI development and deployment can expect economic advantages in manufacturing, services, healthcare, agriculture, and virtually every sector. Conversely, nations falling behind risk economic stagnation as competitors achieve superior efficiency and capability.

Military Applications

AI promises to transform warfare through autonomous weapons systems, enhanced intelligence analysis, cyber capabilities, logistics optimization, and decision support. Military AI leadership could provide decisive advantages in future conflicts. Both nations are investing heavily in military AI, viewing it as critical to maintaining or achieving military superiority.

Surveillance and Social Control

AI enables surveillance at scales and with capabilities previously impossible, like facial recognition tracking individuals through cities, analysis of communications for dissent, predictive systems identifying threats before they materialize. For authoritarian regimes like China, AI provides tools for social control. For democracies, AI raises concerns about privacy and civil liberties even as it offers security benefits.

Technological Standards and Norms

The nation shaping AI development establishes technical standards, ethical frameworks, and governance norms that influence global AI deployment. This affects everything from data protection and privacy to acceptable uses and safety requirements. Standards-setting confers both economic advantages (domestic products aligned with international standards) and soft power (influence over how other nations develop and deploy AI).

Innovation Ecosystems

AI leadership requires ecosystems spanning research institutions, technology companies, capital markets, skilled talent, and supporting infrastructure. Nations with thriving AI ecosystems attract investment, talent, and commercial opportunities. Conversely, falling behind in AI could precipitate broader technology sector decline. The strategic significance means AI competition is zero-sum in some respects. Military advantages or standards-setting influence gained by one nation may disadvantage the other, while positive-sum in others, as AI advances could benefit humanity regardless of which nation achieves them.

Research and Innovation Landscape

Assessing research leadership in AI requires examining both the quantity and quality of research output. By publication volume, China has surpassed the United States in AI research papers. Chinese researchers now publish more AI papers annually than their American counterparts, reflecting the expansion of China's research enterprise. Chinese universities graduate far more STEM PhDs than American institutions, providing a large researcher base.

Publication quantity doesn't necessarily indicate research quality or impact. Citation analysis, measuring how often papers are referenced by subsequent research, suggests American AI research maintains a higher average impact. Papers from leading US institutions like Stanford, MIT, and Carnegie Mellon are cited more frequently than those from most Chinese institutions, indicating greater influence on the field's development. The most highly cited AI papers remain predominantly from American researchers and institutions.

Top-tier AI conference acceptances provide another quality metric. Conferences like NeurIPS, ICML, and CVPR are highly selective, accepting only papers meeting rigorous standards. Analysis of acceptances shows substantial representation from both nations, with the Chinese presence growing. US institutions are still dominant among the most influential papers. Corporate research labs, particularly Google, Microsoft, and OpenAI in the US and Baidu, Alibaba, and Tencent in China, are key contributors to top conferences.

Breakthrough innovations tend to emerge more from US institutions and companies. The Transformer architecture underlying modern large language models came from Google researchers. GPT models and ChatGPT emerged from OpenAI. Reinforcement learning breakthroughs often trace to DeepMind (UK, but Google-owned). While Chinese researchers contribute important advances, foundational innovations most often originate in the American ecosystem.

Several factors explain this pattern. US research culture emphasizes creativity and risk-taking, with greater tolerance for unconventional approaches. American institutions provide autonomy, allowing researchers to pursue novel directions without heavy bureaucratic oversight. The ecosystem surrounding US research institutions (venture capital, entrepreneurial culture, mobility between academia and industry) facilitates translating research into applications. English language dominance in AI research advantages native English speakers and American institutions.

China recognizes the gaps in quality and is working hard to address them. Top Chinese universities are recruiting prominent researchers, improving facilities, and increasing research funding. Chinese technology companies are establishing advanced research labs that can compete with their American counterparts. Government initiatives emphasize moving beyond incremental research toward breakthrough innovation. Progress is evident since Chinese AI research quality has improved substantially, although significant gaps still remain.

Corporate AI Leadership

The world's most valuable and influential AI companies are predominantly American. Google, Microsoft, Amazon, Meta, and Apple lead AI. They possess enormous resources, vast user bases providing data, and the ability to attract top talent. These companies invest billions annually in AI research and development, operate at scales enabling experimentation at unmatched scope, and integrate AI across products that serve billions of users globally.

OpenAI, despite being a relative newcomer, has achieved extraordinary influence through GPT models and ChatGPT. Anthropic, founded by former OpenAI researchers, represents another example of American AI entrepreneurship. NVIDIA's dominance in AI accelerator chips gives the United States critical advantages in compute. American companies like Tesla pioneer autonomous vehicle technology, while others lead in AI applications from drug discovery to robotics.

China's technology giants (Baidu, Alibaba, and Tencent) are formidable competitors with substantial AI capabilities. Baidu leads Chinese search and has invested heavily in autonomous vehicles. Alibaba's cloud computing division provides AI services throughout Asia. Tencent's massive social media and gaming platforms generate data enabling advanced AI applications. Bytedance's TikTok algorithm demonstrates Chinese AI capability in content recommendation. SenseTime and Megvii lead in computer vision and facial recognition.

Chinese AI companies face constraints that limit their global reach. Geopolitical tensions restrict market access to their products in the United States and allied nations. Concerns about security, data privacy, and government relationships create barriers to adoption. Language and cultural factors limit some applications. While Chinese AI companies dominate their domestic market, their international presence remains limited compared to the global reach of American companies.

The regulatory environment differs significantly between the two countries. US companies operate with relatively light regulation, enabling rapid innovation and deployment, though this is changing as concerns about AI risks grow. Chinese companies face tighter government oversight, with requirements to align AI development with government priorities and restrict content deemed sensitive. This control enables coordination but may constrain innovation that requires experimentation beyond the approved boundaries.

American AI companies' access to global talent markets provides advantages as they recruit from worldwide pools, attracting top researchers regardless of nationality. Chinese companies primarily hire Chinese nationals or ethnic Chinese from abroad, limiting their talent base. American companies also benefit from deep integration with the global internet ecosystem, while Chinese companies operate largely within China's internet, bounded by the Great Firewall.

Universites and Research Institutions

American universities remain preeminent in AI research and education. Stanford, MIT, Carnegie Mellon, Berkeley, and other leading institutions produce groundbreaking research and train many of the world's top AI researchers. These universities maintain cultures of academic freedom, interdisciplinary collaboration, and close industry partnerships that facilitate innovation. Faculty move fluidly between academia and industry, with sabbaticals at companies and researchers founding startups based on their university research.

China has invested heavily in expanding AI at their universities. Tsinghua University, Peking University, and other elite Chinese institutions have improved dramatically by recruiting international faculty, establishing well-funded labs, and producing high-quality research. The Chinese government has designated universities for focused AI development with substantial funding. Undergraduate and graduate AI programs have expanded massively, producing large numbers of trained researchers.

Several constraints limit Chinese universities' research impact. Political controls affect academic freedom. Sensitive topics are restricted and ideological conformity is required. This can stifle the often unconventional thinking that is essential for breakthrough innovation. Faculty recruitment focuses heavily on ethnic Chinese researchers, which limits the diversity of perspectives. Movement between academia and industry, while increasing, remains less fluid than in the United States. Bureaucratic processes can slow research and constrain flexibility.

Government research institutions play different roles in each nation. US national laboratories and agencies like DARPA, NSF, and NIH fund substantial AI research, but generally don't conduct research directly (except in specialized areas like nuclear weapons or space). They fund university and corporate research through competitive grants. The Trump Administration's genesis mission provides funding and resources to American national laboratories. China's government research institutes, such as the Chinese Academy of Sciences, conduct substantial research directly with government funding, complementing university research efforts.

Talent and Human Capital

China's massive population provides an enormous potential talent base for AI. The country graduates over 4.5 million STEM bachelor's degrees annually compared to roughly 500,000 in the United States. Chinese students consistently score at the top in international assessments of math and science capabilities. This quantitative advantage in a STEM-educated workforce provides a substantial foundation for AI development.

Once again, quantity doesn't guarantee quality. Critics note that Chinese education emphasizes memorization and test-taking over creativity and critical thinking; skills particularly valuable in cutting-edge AI research. American education's emphasis on problem-solving, interdisciplinary thinking, and intellectual independence may produce fewer graduates, but they cultivate capabilities more valuable for innovation. Evidence on which system better prepares AI researchers remains contested, with strengths and weaknesses in both approaches.

Graduate education presents a complex picture. Chinese universities award more PhDs in AI-related fields than American institutions, and the gap is widening. But many Chinese students pursue PhDs at American universities, with Chinese nationals comprising the largest group of international students in US STEM programs. After graduation, many have in the past remained in the United States rather than returning to China, representing a "brain drain" benefiting American AI development.

This pattern is changing. China's improving research environments, growing salaries, and expanding opportunities make returning home to China an increasingly attractive option. Government programs specifically recruit overseas Chinese researchers with generous packages of pay and benefits. Some Chinese researchers in the United States face visa difficulties, security clearances denials, or suspicion due to geopolitical tensions, making returning home more appealing. The flow of Chinese talent to the United States continues but has slowed, while the reverse flow has increased.

America has significant advantages in attracting global talent. The United States draws researchers from around the world; India, Europe, Latin America, and elsewhere. This diversity brings varied perspectives which are invaluable for innovation. American immigration policies have historically enabled talented foreigners to study, work, and remain permanently in the US. Many leading American AI researchers are either immigrants or children of immigrants. Yann LeCun (France), Chief AI Scientist at Meta, Geoffrey Hinton (UK), the Godfather of Deep Learning, and Fei-Fei Li (China), Professor at Stanford University are a few examples of foreign born researchers that have contributed to American AI.

China's talent attraction focuses primarily on ethnic Chinese researchers from abroad. While overseas Chinese communities include many talented individuals, this approach accesses a smaller global talent pool than the US. China's language barrier, cultural adjustment challenges, and the political environment make attracting non-Chinese talent difficult. Some foreign researchers work in China temporarily, but few settle permanently.

Brain Circulation and Competition

The concept of "brain circulation," talent flowing between nations rather than permanently migrating, characterizes modern AI talent. Researchers maintain connections across borders, collaborate internationally, and may split time between nations. This creates a knowledge diffusion that benefits both countries.

Technology companies enable this circulation of talent through global offices, remote work arrangements, and international research collaborations. A researcher might be employed by an American company but work partly in China, or vice versa. Academic collaborations span borders, with joint research projects, visiting positions, and conference participation creating dense international networks.

China maintains extensive talent recruitment programs, most notably the "Thousand Talents Plan." This plan provides generous compensation and resources to attract overseas researchers. Some US-based researchers participating in these programs failed to disclose required relationships with Chinese institutions, leading to legal prosecution in some cases. The programs have become controversial, with US authorities viewing them as technology transfer mechanisms that threaten national security.

These tensions create challenges for researchers seeking to maintain international collaborations. Scientists face difficult choices between research opportunities and security concerns. The risk is fragmentation of the global AI research community into separate spheres, reducing knowledge sharing and slowing progress. Both nations lose when barriers prevent talent circulation and international collaboration, although security considerations may justify some restrictions.

Data Resources and Advantages

China possesses significant advantages in data collection at scale. With over 1.4 billion citizens and ubiquitous smartphone adoption, China generates enormous volumes of digital data daily. E-commerce platforms, social media, mobile payments, and other digital services create comprehensive digital footprints. The government's emphasis on "smart cities" with sensors, cameras, and connected infrastructure generates additional data streams.

Chinese technology platforms have captive user bases providing data. WeChat functions as an all-purpose platform combining messaging, social media, payments, and services, creating integrated data profiles unavailable in the more fragmented US ecosystem. Dianping (food delivery), Taobao (e-commerce), and other platforms similarly collect rich, behavioral data. The concentration of digital activity on a few large platforms facilitates data aggregation.

The United States has substantial data resources from its technology companies (Google, Meta, Amazon, and others) serving global user bases. American companies pioneered data-driven business models and continue collecting vast amounts of information. In the US, data is more fragmented across platforms and subject to privacy protections limiting comprehensive aggregation and government access.

Privacy regulations affect data availability differently. The European Union's GDPR restricts data collection and use, which affects American companies that operate in Europe. California and other US states have enacted privacy laws, though less strict than GDPR. These regulations, while protecting privacy, constrain the data that is available for AI training.

China's privacy protections are weaker and unevenly enforced. While China has enacted data protection laws in recent years (the Personal Information Protection Law and Data Security Law), implementation prioritizes government access and national security over individual privacy. This enables more permissive data collection and sharing for AI development, particularly for applications supporting government priorities.

The quality and diversity of data matter beyond quantity. American data reflects global diversity from services used worldwide, potentially training AI systems that generalize better across populations. Chinese data predominantly reflects Chinese users, potentially limiting applicability elsewhere. For applications serving Chinese markets, however, Chinese data provides advantages in relevance and cultural appropriateness.

Surveillance Infrastructure and Data Access

China's surveillance state generates extraordinary data that is useful for certain AI applications. Cities are blanketed with cameras equipped with facial recognition that track individuals' movements. Communications are monitored, transactions recorded, and digital behaviors logged. The social credit system integrates data from multiple sources to assess individuals' trustworthiness, thus creating comprehensive profiles of Chinese citizens.

This surveillance infrastructure serves both social control and AI development. The data enables training systems for facial recognition, behavior analysis, anomaly detection, and prediction. Chinese AI companies developing surveillance technologies benefit from access to real-world deployment data at scale unavailable elsewhere. This has made Chinese companies world leaders in surveillance AI, and they export these technologies to governments worldwide.

Democratic countries, including the United States, face ethical and legal constraints limiting surveillance and data collection. American surveillance capabilities are substantial, but domestic surveillance faces constitutional limits and public opposition. Private companies collect data but they must balance commercial interests against privacy concerns and regulatory compliance. This constrains data available for AI training compared to China's more permissive environment.

China's surveillance advantages are domain-specific. For consumer applications in open societies with privacy protections, Chinese surveillance-derived AI offers limited advantages and may face market resistance. Surveillance technologies optimized for authoritarian control may not transfer effectively to democratic contexts that have different norms and legal frameworks.

Data Localization and Cross-Border Flows

Data localization requirements mandating that data about a nation's citizens be stored and processed within that nation affect AI competition. China requires extensive data localization, limiting Chinese data's availability for foreign AI development while at the same time ensuring that Chinese companies and government can access it. The EU's GDPR restricts data transfers outside Europe without adequate protections.

The United States has historically favored free data flows, which benefits American companies that operate globally, as most do. Recently, concerns about foreign access to American data, particularly by the Chinese, have prompted restrictions. For example, TikTok faces pressures to divest or restrict data access due to concerns about the Chinese government's access to American user data.

Data localization fragments the global data ecosystem, creating regional data pools rather than global ones. This can reduce AI system quality if training data isn't representative of diverse populations. It provides advantages to companies operating within regions where they can access data while disadvantaging those operating elsewhere. For US-China competition, data localization reinforces separate ecosystems in each country, with advantages and disadvantages for both.

Strategic competition over data extends to standards-setting for cross-border data flows. The United States promotes frameworks enabling relatively free flows with privacy protections. China advocates sovereignty-based approaches giving nations control over data within borders. These competing visions reflect different governance models and strategic interests, with implications for which approaches other nations adopt.

Computational Infrastructure

Advanced semiconductors, particularly AI accelerator chips like GPUs and specialized AI processors, are critical for training and deploying sophisticated AI systems. This creates a strategic chokepoint in AI competition where the United States holds a significant advantage.

NVIDIA dominates AI accelerator chips with its GPUs, which are essential for training large AI models. The company's CUDA software ecosystem and successive hardware generations have created advantages that competitors struggle to match. NVIDIA is an American company, and US export controls restrict what products can be sold to China. These controls, implemented in 2022 and tightened in 2023, prohibit exporting the most advanced AI chips to China, which constrains Chinese AI development.

Advanced semiconductor manufacturing is dominated by Taiwan Semiconductor Manufacturing Company (TSMC), which produces cutting-edge chips at a scale and quality no other manufacturer matches. While TSMC is Taiwanese rather than American, the US alliance with Taiwan and its influence over TSMC's operations provide indirect control. TSMC relies on American chip design tools from companies like Synopsys and Cadence, and lithography equipment from the Netherlands (ASML), over which the US exerts also influence. This manufacturing edge gives the United States leverage over semiconductor supply chains.

China recognizes this vulnerability and has invested heavily in developing domestic semiconductor manufacturing. Substantial government funding supports both Chinese chip manufacturers like SMIC and foundries that develop advanced manufacturing products. But China still lags behind the cutting edge in manufacturing technology, by at least two generations. Efforts to acquire or develop advanced lithography have struggled, with ASML unable to sell its most advanced EUV machines to China due to Dutch government restrictions, imposed under US pressure.

Chinese companies have developed AI accelerator chips, including Huawei's Ascend processors and chips from startups like Biren and Cambricon. These chips provide capabilities for Chinese AI development in spite of the export controls. But they generally trail NVIDIA's latest offerings in performance and the maturity of the software ecosystem. Chinese cloud providers offer access to AI computing, but the most demanding applications, like training frontier AI models, are constrained by chip performance limitations.

The semiconductor situation creates asymmetric interdependence. American AI development benefits from unfettered access to the most advanced chips, while Chinese development faces constraints. The global semiconductor supply chain is complex, with American chips manufactured in Taiwan using equipment from the Netherlands, Japan, and elsewhere. Disruptions to these supply chains, from conflict, natural disasters, or political decisions, could harm American capabilities. The concentration of advanced manufacturing in Taiwan creates vulnerabilities for both nations, particularly given the tensions over Taiwan's status, although TSMC has recently opened a semiconductor manufacturing plant in the US.

Supercomputing and Cloud Infrastructure

Supercomputing capability, as measured by floating-point operations per second and rankings like the TOP500 list, provides infrastructure for AI development. China historically has competed strongly in supercomputing, with systems like Sunway TaihuLight and Tianhe-2 topping the rankings. in recent years, China stopped submitting systems to international rankings which creaties uncertainty about their current capabilities. US export controls on high-end processors constrain Chinese supercomputer development, although China has developed indigenous processors for supercomputing.

For AI, cloud computing infrastructure matters more than traditional supercomputers. As opposed to traditional supercomputing architectures, training large models requires coordinated clusters of AI accelerators with high-bandwidth interconnects. American companies (AWS, Microsoft Azure, Google Cloud) dominate global cloud computing and provide leading AI training infrastructure. These cloud providers operate data centers worldwide with massive AI accelerator installations.

Chinese cloud providers (Alibaba Cloud, Tencent Cloud, Huawei Cloud) serve primarily the Chinese market and have expanded regionally in Asia. They provide substantial AI computing domestically, but they lack the global reach and scale of American hyperscalers. Export controls limiting advanced chip access constrain their ability to offer cutting-edge AI training, though they can serve many applications with available hardware.

The concentration of AI training capability in major cloud providers creates dependencies for smaller companies and researchers. American dominance in cloud computing provides economic benefits and strategic influence, controlling the infrastructure on which others depend. But it also creates concerns from centralization and potential vulnerabilities if cloud providers become targets for attack or disruption.

Energy availability affects computational infrastructure development. AI training consumes enormous amounts of electricity. Training large models can require energy measured in megawatt-hours. Both nations have substantial electricity generation capacity. The United States has diverse sources including natural gas, renewables, coal, and nuclear. China relies heavily on coal, though they are rapidly expanding renewables. Electricity costs and availability can influence where AI data centers locate, with both nations offering competitive options.

Software Ecosystems and Frameworks

While hardware enables AI, software frameworks and tools determine the ease of development. American companies have created dominant AI frameworks like TensorFlow (Google), PyTorch (Meta), and others. These open-source frameworks are used globally, including extensively in China. This gives American companies an influence over AI development practices and standards. Developers learn tools, create compatible code, and share models using these frameworks.

China has developed indigenous AI frameworks. These include Baidu's PaddlePaddle, Huawei's MindSpore, and others. These serve domestic markets and provide alternatives if American frameworks become unavailable due to sanctions or restrictions. These tools have far smaller user bases and ecosystems than TensorFlow or PyTorch, limiting their influence.

The open-source nature of leading frameworks complicates the competitive atmosphere. While American companies created them, anyone globally can freely use them. This enables Chinese AI development, though it also helps American dominance by ensuring developers worldwide use American-created tools. Efforts to restrict access to open-source AI tools would face a backlash. Restrictions on cutting-edge proprietary systems are more feasible.

Software stacks extending beyond frameworks to development tools, model repositories, deployment platforms, and MLOps infrastructure are dominated by American companies and open-source projects. GitHub (Microsoft-owned), Hugging Face, and others provide essential infrastructure for AI development, and are used globally. This gives American platforms visibility into AI development trends and influence over practices.

Commercial Applications and Economic Impact

AI adoption across industries is occurring rapidly in both nations, but with different patterns and priorities. The United States leads in AI integration in technology, finance, healthcare, and professional services. American companies pioneered AI-powered recommendations, search, advertising, and cloud services. Financial services firms use AI for algorithmic trading, fraud detection, risk assessment, and customer service. Healthcare applications include diagnostics, drug discovery, and personalized medicine.

China excels at integrating AI into manufacturing, logistics, retail, and the urban infrastructure. Chinese manufacturers use AI for quality control, predictive maintenance, and process optimization at scales enabled by China's hugh manufacturing base. Logistics companies like JD.com and Cainiao employ AI extensively for warehouse automation, delivery optimization, and supply chain management. Retail AI applications include cashierless stores, personalized commerce, and inventory management. "Smart city" initiatives integrate AI into traffic management, public safety, and urban services.

The economic impact of AI adoption is substantial in both countries. Experts estimate AI could contribute trillions to global GDP in the coming decades, with the US and China as primary beneficiaries. US strengths in high-value services and technology sectors position it well to capture the economic benefits of AI. China's manufacturing prowess and large domestic market provide advantages for AI-driven industrial and consumer applications.

The effects of automation's employment differ between the nations. The United States faces concerns about AI-driven unemployment in industries from trucking to customer service. China confronts similar challenges, but with different priorities. The government prioritizes economic development and global competitiveness over employment preservation, though it maintains vigilant about social stability if unemployment rates rise.

Autonomous Vehicles and Transportation

Autonomous vehicle development represents a prominent AI battleground. American companies like Tesla, Waymo (Google), and others have invested heavily in self-driving technology. Tesla's Autopilot and Full Self-Driving systems are deployed in millions of vehicles, accumulating vast real-world driving data. Waymo operates commercial robotaxi services in several cities. American companies generally lead in autonomous vehicle technology, though no system yet achieves full autonomy in all conditions.

Chinese companies including Baidu, Pony.ai, WeRide, and AutoX are developing autonomous vehicles with substantial government support. China has designated cities for autonomous vehicle testing. They have created controlled environments accelerating development. Chinese companies benefit from government coordination with regulations enabling testing, infrastructure supporting autonomous vehicles, and favorable policies encouraging adoption. Baidu's Apollo platform aims to become the standard infrastructure for autonomous driving.

Regulatory approaches differ between the two countries. US regulation is fragmented across federal agencies and state governments, creating inconsistent requirements that complicate nationwide deployment. States like California require extensive reporting and testing, while others take a hands-off approach. This fragmentation slows deployment.

China's centralized regulatory structure enables faster policy development and deployment. The government can designate regions for autonomous vehicle testing, mandate infrastructure support, and coordinate across agencies. This accelerates deployment, though centralized control also means mistakes can affect the entire country. Safety validation processes differ, with China potentially accepting higher risks to achieve faster deployment.

The massive Chinese domestic market provides advantages for scaling autonomous vehicle deployment. If Chinese companies achieve commercially viable autonomy, they have enormous domestic opportunities. American companies face strong domestic and European demand as well as substantial competition. International expansion for autonomous vehicles faces unique challenges given the requirements for local testing and regulatory approval in each market.

Facial Recognition and Surveillance

China has the global lead in facial recognition and surveillance AI. Advanced systems have been deployed for public security, access control, and identification. Chinese companies including SenseTime, Megvii, and Yitu have developed sophisticated facial recognition applications with accuracy exceeding those of humans. These systems are integrated into public security infrastructure, enabling such uses as finding missing persons and tracking dissidents.

The sheer scale of deployment provides Chinese companies with advantages in training data and real-world testing. Surveillance cameras throughout Chinese cities generate billions of images for training recognition systems. Public acceptance and weak privacy protections enable deployment in contexts that would face opposition elsewhere. This has made Chinese facial recognition technology among the world's most advanced.

Internationally, these technologies face ethical concerns and market limitations. Democratic societies impose restrictions on mass surveillance, and limit the deployment of Chinese-style systems. Facial recognition has faced bans or moratoria in various US jurisdictions due to concerns about privacy and civil liberties. American companies developing facial recognition, including Amazon, Microsoft, and others, have faced substantial criticism and in some cases paused projects and sales to law enforcement.

American AI capabilities in facial recognition and surveillance are substantial, and deployed selectively. Intelligence and law enforcement agencies use advanced systems, but domestic deployment faces legal and political constraints. American companies lead in some related areas like satellite imagery analysis and open-source intelligence, with applications in military intelligence and commercial services.

The surveillance AI competition raises fundamental questions about values and governance. China views surveillance as enabling security and social stability, with individual privacy subordinate to collective goals. Western democracies emphasize privacy rights and place limits on government surveillance. This difference in values means AI applications considered acceptable in China face rejection elsewhere.

Healthcare AI Applications

Healthcare represents an area where AI promises transformative benefits, and both nations are investing heavily in AI to make it happen. American strengths include leading healthcare research institutions, pharmaceutical giants, and medical device companies that are pioneering AI applications. FDA-approved AI medical devices now include systems for detecting diabetic retinopathy, analyzing medical images, predicting sepsis, and many other applications.

AI drug discovery has attracted massive investment in the United States, as companies use machine learning to identify drug candidates, predict molecular properties, and optimize clinical trials. American biotechnology firms lead in applying AI to genomics, personalized medicine, and biological research. The close integration of American universities, hospitals, and medical tech companies facilitates translating research into clinical applications.

China has advantages in population scale for health data, genomic databases, and clinical trial recruitment. Chinese hospitals generate enormous patient data volumes, and genome sequencing initiatives have collected genetic information from millions. This data enables training AI systems for disease diagnosis, treatment optimization, and drug discovery specific to Chinese populations.

Healthcare AI faces regulatory, ethical, and practical challenges in both nations. Medical AI must meet safety standards and clinical validation requirements, processes that are lengthy and rigorous. US regulations from the FDA provide clear frameworks with slow approval. China's regulatory processes for medical AI are evolving, potentially enabling faster deployment but with questions about safety validation.

Privacy concerns affect healthcare AI differently in each nation. American healthcare data is protected by HIPAA regulations restricting use without patient consent. Although HIPAA protects privacy, it complicates AI development that requires large datasets. China's weaker privacy protections enable more permissive health data use and accelerates AI development though its practices raise ethical concerns regarding consent and data protection.

The healthcare systems differ fundamentally, with American healthcare being fragmented and largely private, while China's is more centralized and government-dominated. This situation affects AI deployment patterns. Chinese healthcare AI might achieve more uniform deployment through government mandates, while American deployment is market-driven and highly regulated. Neither approach is clearly superior. Chinese centralization enables coordination but risks errors affecting millions, while American fragmentation enables experimentation but creates inequities in accessibility.

Military and Intelligence Applications

The United States and China approach military AI with different doctrines reflecting their military traditions and strategic situations. US military doctrine emphasizes technological superiority, precision strike, network-centric warfare, and minimizing casualties. AI fits naturally into this doctrine, promising enhanced intelligence, improved targeting, autonomous systems, and decision support enabling information advantages over adversaries.

American military services are pursuing AI applications in many areas: autonomous drones, underwater vehicles, AI-enhanced cyber operations, intelligence analysis systems, predictive maintenance for equipment, and logistics. The Department of War established the Joint Artificial Intelligence Center (JAIC) to coordinate military AI development, and it has made AI a priority across the service branches. Concepts like "mosaic warfare" envision AI’s ability to coordinate numerous autonomous systems in use in adaptive networks.

China's military modernization explicitly includes AI as a transformative technology. Chinese military writings discuss "intelligentized warfare" as succeeding "informatized warfare." This positions AI as enabling new military capabilities and operational concepts. The People's Liberation Army is developing autonomous systems, AI-enhanced command and control, intelligent logistics, and military applications of civilian AI technologies through civil-military fusion strategies.

Chinese military AI development benefits from close integration between civilian technology companies and military institutions. Major Chinese tech firms explicitly support military modernization, with less separation between commercial and military sectors than in the United States. This "civil-military fusion" strategy enables rapid military adoption of civilian AI advances, which allows rapid transitions of technology from research to deployment.

The US military-industrial complex maintains stronger boundaries between commercial technology and military applications. Major defense contractors like Lockheed Martin, Raytheon, and Northrop Grumman develop military AI. Some Silicon Valley companies have resisted military contracts due to employee opposition. For example, Google notably canceled its Project Maven contract after internal protests, although others including Microsoft, Amazon, and Palantir actively pursue defense work.

Ethical and legal frameworks differ significantly between the two countries. US military operations are constrained by laws of armed conflict, strict rules of engagement, and political accountability. Public controversy over autonomous weapons and "killer robots" has prompted policy debates about the need for human control. The Department of War has adopted principles of AI ethics that emphasize responsible development.

China's authoritarian system faces fewer constraints from public opinion or independent oversight. While China officially endorses international humanitarian law, its military faces less transparency and accountability than US forces. Chinese writings on military AI emphasize winning future conflicts and achieving technological surprise, with less visible debate about any ethical constraints on AI weapons systems.

Autonomous Weapons and Drone Technology

Autonomous weapons systems that can select and engage targets without human intervention represent perhaps the most controversial military AI application. Both nations are developing increasingly autonomous systems, with differing policies and philosophies regarding human control over lethal decisions.

The United States leads in military drone technology, with extensive operational experience from decades of armed drone operations. American drones like the MQ-9 Reaper are remotely piloted rather than autonomous, but they increasingly incorporate AI for functions like target recognition, flight control, and sensor analysis. The military is developing more autonomous systems including swarming drones that coordinate without human control and loitering munitions that autonomously search for targets.

US policy officially requires "appropriate levels of human judgment" over lethal force decisions. The military distinguishes between systems that are autonomous in some functions (navigation, target detection) and human-controlled for lethal functions, from fully autonomous weapons that may independently decide to kill. Current policy prohibits the latter, though technology is advancing toward such capabilities.

China has developed extensive drone capabilities, with systems ranging from small tactical drones to large-combat, unmanned aerial vehicles. Chinese drones have been exported widely, particularly to countries unable or unwilling to purchase American systems. Some Chinese drones reportedly include autonomous targeting capabilities, though operational use and human-in-the-loop requirements are opaque. China has not articulated detailed public policy on autonomous weapons as compared to US policy.

The proliferation of autonomous weapons technology creates risks neither nation fully controls. As AI-enabled weapons spread globally, including to non-state actors, the potential for accidents, escalation, or unauthorized use increases. Both nations have incentives to develop autonomous weapons. They promise military advantages and neither wants to fall behind the other. This condition creates a security dilemma where mutual restraint is difficult to achieve, a harbinger to the Cold War dynamics between nuclear powers.

International efforts to regulate autonomous weapons through the United Nations Convention on Certain Conventional Weapons have made limited progress. Some nations advocate bans on autonomous weapons, while major powers including the US and China resist legally binding restrictions on their military AI development. The lack of consensus enables continued development with limited international constraints.

Cyber Operations and Information Warfare

AI is transforming cyber operations and information warfare, areas where the US-China competition is intense. Machine learning enhances both cyber offense and defense, enabling more sophisticated attacks, better intrusion detection, automated response, and the rapid analysis of network traffic for anomalies.

The United States possesses advanced cyber capabilities, developed by the NSA, US Cyber Command, and other agencies. American cyber operations have reportedly disrupted adversarial networks, collected intelligence, and defended critical infrastructure. AI enhances these operations through automated vulnerability discovery, adaptive attack strategies, and the processing of vast amounts of network data to identify threats. Private sector cybersecurity companies supplement government work with cutting-edge detection and response systems.

China's cyber capabilities have grown substantially, with extensive operations attributed to Chinese government-linked groups. Chinese cyber espionage has targeted US government networks, defense contractors, technology companies, and critical infrastructure. AI enables more sophisticated operations like spear phishing using AI-generated personalized messages, malware that adapts to evade detection, and the automated analysis of stolen data for valuable intelligence.

The attribution challenge - determining the responsibility for cyber attacks - complicates responses. AI enables more sophisticated false flag operations and obfuscation of attack origins. Both nations conduct cyber operations through proxies, contractors, and techniques that create plausible deniability. This ambiguity makes deterrence difficult and increases risks of escalation.

Information warfare, defined as influencing populations through propaganda, disinformation, and narrative manipulation, is enhanced by AI. Large language models can generate convincing fake content at scale. Deepfakes create realistic but fabricated videos, and algorithms identify optimal targeting for influence campaigns. Both nations use these techniques, though with different objectives and different constraints.

Chinese information operations control narratives about China, promoting authoritarian governance models, and undermining confidence in democracies. Operations target Chinese diaspora populations, suppress criticism of Chinese government, and promote narratives favorable to Chinese interests. AI enables scaling these operations and personalizing messaging to affected audiences.

US information operations officially focus on countering adversary disinformation and promoting American values. American social media platforms are vehicles for both organic communication and influence operations. AI content moderation attempts to limit malicious information while preserving free expression.

The weaponization of AI for information warfare threatens democratic discourse and social cohesion. When citizens cannot trust information sources and authentic content is indistinguishable from AI-generated fakes, informed democratic deliberation becomes difficult. Authoritarian systems may be less vulnerable to some information warfare tactics, though digital technologies create channels beyond government control, even in China.

Intelligence Collection and Analysis

Intelligence agencies in both nations use AI for collection, processing, and analysis. The volume of intelligence data (satellite imagery, communications intercepts, open-source information, human intelligence reports) overwhelms human analysts. AI enables processing this flood of information to identify patterns, anomalies, and actionable intelligence.

US intelligence agencies including the CIA, NSA, and NGA have invested heavily in AI. Computer vision analyzes satellite and drone imagery at scales impossible to do manually. Natural language processing translates and analyzes foreign communications. Machine learning identifies patterns in terrorist networks, weapons proliferation, and adversary activities. The intelligence community has established organizations like the National Security Commission on AI to coordinate AI adoption across Federal agencies.

China's intelligence apparatus similarly employs AI extensively. Facial recognition and surveillance systems enable tracking individuals throughout cities. Analysis of digital communications, financial transactions, and social media identifies security threats to the regime. AI processes intelligence from cyber operations, and analyzes stolen data for valuable information. The integration of intelligence, public security, and commercial surveillance creates comprehensive monitoring capabilities.

AI's intelligence applications raise civil liberties concerns, particularly in democracies. The capabilities enabling intelligence collection against foreign adversaries can be turned domestically, threatening privacy and political freedom. US intelligence agencies face legal restrictions on domestic surveillance. China's intelligence and security apparatus faces minimal constraints on domestic surveillance, enabling comprehensive monitoring that would violate constitutional protections in the United States.

The dual-use nature of intelligence AI means that systems developed for national security inevitably have commercial and domestic applications. Facial recognition for border security can be used for domestic surveillance. Natural language processing for foreign intelligence analysis can monitor domestic communications. This situation creates tensions between security imperatives and civil liberties that democracies must navigate.

Industrial Policy and Investment

China's approach to AI development involves extensive government coordination and funding reflecting its state-led economic model. The 2017 AI Development Plan committed substantial resources and established aggressive goals for AI leadership. Provincial and municipal governments launched their own AI initiatives with funding, tax incentives, and subsidized facilities. State-owned enterprises invested in AI as part of a government-directed industrial policy. The coordination between national strategy, government funding, and corporate execution gives China advantages in mobilizing resources toward AI development.

Estimates of Chinese government AI investment vary widely due to lack of transparency, but analysts suggest tens of billions of dollars annually when including national programs, provincial initiatives, subsidies, and state-enterprise investment. This funding supports research institutions, technology companies, infrastructure development, and industrial AI adoption. The government coordinates across sectors, reducing duplication, and aligning efforts toward strategic priorities.

The United States historically has relied more on private sector innovation with government playing a supporting role through research funding. Federal AI research funding through agencies like NSF, DARPA, DOE, and NIH totals several billion dollars annually, a substantial amount but way smaller than overall private investment.

American AI development benefits from the world's most sophisticated venture capital ecosystem, which has invested hundreds of billions in AI startups. This private capital dwarfs government funding and enables rapid experimentation with diverse approaches. The US system allows market-driven resource allocation rather than government direction, with advantages in efficiency and innovation though lacking in coordination with national strategic objectives.

Strategic Industries and National Champions

For government support, China identifies AI, semiconductors, biotechnology, and clean energy as strategic industries. The "Made in China 2025" plan (later de-emphasized due to international criticism) targeted Chinese leadership in advanced technologies. AI features prominently in these plans, with the government identifying and supporting "national champion" companies that are expected to lead on a global scale.

Companies like Baidu, Alibaba, and Tencent receive government support, including favorable regulations, data access, protection from foreign competition, and pressure on state enterprises to purchase their services. SenseTime and other AI companies have received substantial government-backed investments. This approach creates well-resourced companies able to make long-term investments. The approach also risks favoring politically connected firms over more innovative competitors.

The United States lacks a comparable industrial policy, although government procurement, defense spending, and research funding support strategic industries. Tech giants like Google, Microsoft, and Amazon have become powerful largely through market success rather than government designation. They benefit from favorable regulations, government contracts, and research collaborations.

Debates continue in the United States about whether more active industrial policy is needed to compete with China's coordinated approach. Proponents argue that China's model enables rapid progress while the US risks falling behind without comparable coordination. Skeptics contend that market-driven innovation outperforms government direction and that industrial policy invites rent-seeking and political capture. The tension reflects broader American ambivalence about government economic intervention.

Under the Trump Administration, the US has used tariffs to reshore manufacturing, reduce foreign dependence, and strengthen bargaining power in strategic sectors such as pharmaceuticals and semiconductors.

Intellectual Property and Technology Transfer

Intellectual property protection and technology transfer are contentious issues in the US-China AI competition. The United States has accused China of systematically acquiring American technology through cyber espionage. And forced technology transfer, talent recruitment programs, and academic collaboration enables knowledge flows without adequate protections.

High-profile cases of Chinese nationals accused of stealing AI-related technology from American companies, including self-driving car technology from Waymo and trade secrets from various tech firms, have led to prosecutions and intensified scrutiny. The US government has implemented export controls, investment restrictions through CFIUS review of Chinese acquisitions, and university research security requirements to limit technology transfer.

China maintains that it respects intellectual property and that the accusations reflect protectionist attempts to limit Chinese tech progress. China has strengthened IP laws in recent years and increased enforcement, though implementation remains uneven. Chinese officials argue that much of the Chinese technological progress results from indigenous innovation rather than theft, and that American accusations underestimate Chinese capabilities.

The reality likely involves both legitimate concerns about technology acquisition through problematic means and American overreaction that attributes all Chinese progress to theft rather than recognizing genuine innovation. Chinese companies and researchers do conduct significant indigenous research and development. However, evidence of industrial espionage, forced technology transfer, and opaque talent recruitment programs provides grounds for concern.

Technology transfer policies affect the AI competition asymmetrically. American restrictions limit Chinese access to cutting-edge AI technology, particularly semiconductors. The global and networked nature of AI research makes completely preventing technology diffusion difficult to accomplish. Publications, open-source software, international conferences, and talent mobility ensure knowledge spreads despite restrictions. The question is whether transfer restrictions slow Chinese progress meaningfully or primarily harm American interests by reducing access to Chinese markets and collaborators.

Talent Competition and Retention

Both nations compete aggressively for AI talent, but their strategies are very different. The United States traditionally attracts global talent through prestigious universities, leading companies, academic freedom, and immigration opportunities. This remains a significant advantage over China. America's ability to draw researchers from worldwide talent pools provides diversity and scale that China cannot match.

This advantage is eroding due to immigration restrictions, visa processing delays, and geopolitical tensions making the United States less welcoming to international students and researchers. Some Chinese researchers face heightened scrutiny, career obstacles, or choose to return to China rather than face uncertain prospects in America. The loss of even a fraction of Chinese AI talent significantly impacts American capabilities given Chinese nationals' prominence in US AI research.

China's talent recruitment focuses on repatriating ethnic Chinese researchers from abroad with generous compensation, research funding, and career opportunities. Programs target both established researchers and recent graduates. China's improving research environments, emerging tech sector, and opportunities to work on cutting-edge problems make returning home increasingly attractive. Some researchers maintain affiliations in both countries, though this is becoming more difficult as tensions require choosing sides.

Retention of domestic talent challenges both nations. American tech companies compete intensely for AI researchers, offering salaries that universities cannot match. This drains academic departments and concentrates talent in industry, with implications for fundamental research and education. China faces similar dynamics as its tech companies offer compensation rivaling international competitors to retain talent who might otherwise emigrate.

The global competition for AI talent is zero-sum in some respects - researchers working in one nation generally aren't simultaneously contributing to the other - though knowledge diffusion through publications and open-source contributions creates positive spillovers. Both nations benefit from expanding the global talent pool through education, but each nation seeks advantages in attracting and retaining the best researchers.

International Influence and Standards

Both nations seek to expand international influence through AI technology exports and partnerships, with different approaches and advantages. American AI companies have global reach through cloud services, software platforms, and commercial products. Google, Microsoft, Amazon, and others operate internationally, providing AI services to customers worldwide. This creates economic benefits, establishes American technologies as standards, and builds relationships that generate influence.

China exports AI technologies particularly to developing nations through Belt and Road Initiative (BRI) partnerships, development assistance, and commercial sales. Chinese surveillance systems have been exported to dozens of countries in Asia, Africa, Latin America, and elsewhere. Some exports include training and technical assistance, thereby embedding Chinese technologies in partner nations' infrastructure. China markets AI as enabling "leapfrog development" - adopting advanced technologies without building intervening capabilities - which particularly appeals to developing nations.

Chinese AI exports face concerns about surveillance, human rights, and dependency. Critics argue that Chinese surveillance systems enable authoritarian control and that recipient nations become dependent on Chinese technology and expertise. Some systems reportedly include backdoors enabling Chinese access. Countries adopting Chinese AI standards and systems may find themselves in China's technological sphere of influence.

American AI exports face different concerns. Some nations worry about American surveillance and data access, concerns validated by revelations of NSA activities. American products sometimes include capabilities for government access that trouble some foreign partners. High costs of American technology create barriers for developing nations, leaving openings for cheaper Chinese alternatives.

The competition extends to establishing AI partnerships and research collaborations. The United States has AI cooperation agreements with allies (Japan, Australia, and European nations) coordinating research and standard-setting. China has established AI research collaborations throughout Asia, Africa, and Latin America, often combined with infrastructure investment and development assistance. These partnerships create networks of influence and establish norms about AI development and use.

Standards-Setting and Governance Frameworks

International standards for AI are being contested covering everything from technical protocols to ethical principles to safety requirements. The United States promotes frameworks emphasizing innovation, private sector leadership, minimal regulation, and values like transparency and individual rights. This reflects both American political traditions and its advantages in commercial AI.

China advocates frameworks emphasizing sovereignty, government coordination, collective development, and social stability. Chinese proposals stress national control over data and AI development, government roles in directing AI applications, and framing AI governance as state prerogatives rather than issues for international coordination. This reflects the Chinese political system and the desire for influence over global AI norms.

The European Union has emerged as a third force in AI governance with its AI Act and stiff regulatory approach. The EU emphasizes precaution, human rights protection, democratic accountability, and risk-based regulation. Europe's regulatory power, the "Brussels Effect," influences global standards as companies worldwide adjust to comply with EU rules in order to access its market.

Standard-setting occurs through international bodies like the International Organization for Standardization (ISO), International Telecommunication Union (ITU), and various multi-stakeholder forums. The United States and China compete for leadership in these organizations, with different strategies. American delegations typically include private sector participation and emphasize consensus and technical merit. Chinese delegations follow government direction and pursue leadership positions in international organizations to influence standards from within.

The fragmentation of global AI governance with different standards in different regions creates inefficiencies, but also reflects a genuine disagreement about values and priorities. Authoritarian nations prefer governance frameworks that enable government control, while democracies emphasize individual rights and private initiative. Bridging these differences through unified global standards may be impossible, suggesting a future of competing AI governance regimes.

Soft Power and Narrative Competition

Beyond technology and standards, the US-China AI competition involves soft power and competing narratives about AI's role in society. The United States promotes AI's potential for innovation, economic growth, solving global challenges, and improving lives while at the same time requiring guardrails to protect human rights and personal safety. American messaging stresses democratic values, market innovation, and international cooperation.

China promotes narratives positioning AI as a tool for development, social harmony, and effective governance. Chinese messaging emphasizes AI enabling poverty reduction, public services improvement, and social stability. The authoritarian governance model is presented as enabling coordinated AI development without Western inefficiencies. China argues its approach delivers results efficiently while respecting different cultural values about privacy and authority.

These competing narratives appeal to different audiences. Developed democracies generally prefer American approaches that emphasize rights and market innovation. Developing nations and autocratic governments often find Chinese messaging more appealing. The competition for global opinion about appropriate AI development and governance models reflects a broader ideological competition between democratic and authoritarian systems.

Media and information ecosystems shape these narratives. American media companies, English language dominance, and the global reach of American cultural products provide advantages in spreading US perspectives. Chinese media expansion through CGTN and other outlets, Confucius Institutes, and digital platforms provide channels for Chinese narratives, though with less global penetration than American media.