The AI Race With China and the Uncertain Future of Truth
Artificial intelligence (AI) development is at the center of geopolitical competition as both the United States and China seek AI supremacy. The United States has traditionally been the leader in AI research and innovation, but China’s targeted investments and strategic efforts are quickly diminishing this lead. The apparent success of Chinese AI lab DeepSeek in creating advanced models at reduced costs compared to U.S. competitors has raised questions about who will lead AI development in the future.
The competition for AI supremacy, along with advancements in other areas like quantum, will have broad implications as China seeks to surpass the United States in all domains to establish itself as the preeminent global power. The United States and other democracies need to understand how China’s AI dominance will create new technological, economic and military challenges, as well as novel geopolitical threats, like the proliferation of misinformation during regional or global conflicts that challenge even the basic understanding of truth.
DeepSeek: A Disruptive Force in the AI Landscape
The Chinese AI research lab DeepSeek, established by the quant hedge fund High-Flyer, has become a key player in AI research and development. DeepSeek was created during U.S. sanctions that blocked access to AI chips, demonstrating China’s commitment to AI development despite external restrictions. DeepSeek is different from other commercial AI companies because it has chosen to open-source its models, which lets anyone use and develop its technology further.
U.S. companies like OpenAI and Google use closed-source AI models; open-sourcing is disruptive and enables more open collaboration and innovation throughout the AI sector. DeepSeek’s stated focus remains on research without detailed commercialization plans, which has enabled its technology to avoid China’s strictest AI regulations. The Chinese national priority on aggressively funding its AI ecosystem has allowed startups such as DeepSeek to compete globally.
DeepSeek’s AI models perform remarkably well in reasoning, math and coding tasks. The R1 model demonstrates performance levels comparable to OpenAI’s modern models but comes with significantly lower training and deployment costs. DeepSeek’s breakthrough shocked the AI industry and sparked concerns about America’s leadership status. These worries cast doubt on major AI investments and market valuations in the sector, resulting in a $600 billion paper loss for NVIDIA.
DeepSeek succeeded through its unique model architecture and efficient training methods while specializing in domains such as mathematics and coding. DeepSeek achieved high performance through task-specific model optimization and reinforcement learning techniques while using fewer computational resources. DeepSeek’s R1 model training was said to cost $5.6 million—a fraction of what similar models from U.S. companies require for training.
However, the Financial Times found that DeepSeek had actually invested more than $500 million in NVIDIA GPUs while skirting sanctions and applied model distillation to OpenAI’s GPT-4, which allowed it to transfer its knowledge and obtain a significant advantage during training. In some conversations, DeepSeek-R1 even insisted it was ChatGPT.
DeepSeek’s success also raises security concerns. DeepSeek, like TikTok, collects extensive information from user interactions, including all chats, which it stores on servers located in China. Data security and privacy concerns are even more acute due to China’s national security laws, which force companies to assist its intelligence services. Research has also shown that DeepSeek sends user data to ByteDance, TikTok’s parent company, and includes code that sends user identity information to China Mobile, a state-operated telecom company.
But more troubling is DeepSeek’s censorship. DeepSeek presents itself as an open platform but adopts Chinese Communist Party (CCP) information control policies. The platform intentionally omits any mention of the Tiananmen Square massacre and redirects questions about it to less “controversial” subjects, like games or coding. DeepSeek parrots CCP positions when users inquire about “sensitive” topics such as Taiwan, Tibet, Xinjiang, Hong Kong and the South China Sea. This raises serious concerns about misinformation spreading among students and younger users globally, and in the broader public discourse.
The CCP uses DeepSeek as a narrative control and global perception-shaping tool, similar to its targeted use of TikTok. The rapid adoption of DeepSeek by American and Western youth parallels TikTok’s broad reach and creates major concerns about its ability to propagate CCP propaganda and skew perceptions of key historical and political realities. Note that this doesn’t only apply to the DeepSeek app or website: the responses are the same even from models downloaded and run locally.
China’s AI Strategy: Global Dominance
The Chinese government has launched its AI development through research and development investments, along with the creation of a strong AI ecosystem and national AI research centers. The Chinese government has emphasized AI education and talent cultivation to build a capable workforce necessary for reaching its AI objectives. China’s “New Generation AI Development Plan” reflects its strategic focus on near- and long-term priorities to secure dominance in global AI markets, technologies and systems by 2030.
The governance model China utilizes for AI development is much different from the approaches adopted by the United States and the European Union. The United States has embraced market-driven AI development while the EU emphasizes ethical AI standards and China maintains strict AI supervision to ensure alignment with national strategic goals. The country’s AI governance encompasses data localization rules, algorithm registration requirements and content moderation controls. China also targets traditional industries with AI applications to push its value chain advancement strategy.
Through open-source AI initiatives, China also drives innovation and talent acquisition while positioning itself to dominate global AI development by disrupting the current commercial market dynamics. China seeks to establish itself as an AI leader and shape future AI development by making its AI models and technologies broadly accessible. This strategy is intended to destabilize the existing commercial AI sector.
DeepSeek has shown that advanced performance can be attained with limited resources, regardless of the specific technologies and circumstances with training R1. This development has significant implications for the AI landscape. It levels the playing field, potentially allowing smaller companies and countries to more effectively enter the AI space and compete.
Geopolitical Implications: New Areas of Competition
The AI competition between the United States and China has major geopolitical ramifications. AI is more than just a technology—it is a powerful instrument to boost economic productivity and military and intelligence capabilities and to shape global norms and standards.
China’s strategy to attain AI superiority intends to diminish U.S. influence. AI advancements could give China critical advantages across diverse domains, from manufacturing to finance, and defense sectors. If the global balance of power in the AI space shifts toward China, it could lead to rising tensions, instability and potential conflict.
As the world is fractured into separate regulatory and technological domains, there is a risk of escalating tension between the United States and China. This environment raises questions about the interoperability of AI systems, and the risk of a new “digital divide” emerging between nations with varying AI technologies and standards—with some choosing U.S. technologies and others choosing China’s AI “Belt and Road.”
The widespread adoption of generative AI creates geopolitical challenges for both the United States and its allies. The race to lead in AI increases competition between countries, as governments seek to use AI for economic gain and technological leadership and to influence international standards and norms. Deepfake videos, highly targeted disinformation campaigns and cyber attacks generated by AI systems at scale all present dynamic geopolitical risks that threaten national and global security and trust in democratic institutions.
Conclusion: Navigating AI Uncertainty
The DeepSeek affair has significant implications for the United States. It shows that massive computing and data center investment does not guarantee AI dominance and highlights the need for strategic adaptability and agility in AI sector competition. To maintain its leadership position, the United States needs to prioritize innovation along with open-source AI development and enhancement of its own AI ecosystem.
China’s strategic approach toward AI development has the potential to remake global power structures. If China achieves its goal of becoming the world’s leading AI power, it will have decisive advantages across multiple sectors. The United States and its allies must unite to challenge China’s AI ambitions and promote ethical and responsible AI use worldwide.
Addressing AI-enabled misinformation and cyber attacks requires a multipronged strategy. The approach needs frameworks for ethical AI governance development and investments in AI safety while promoting media literacy and critical thinking skills. International cooperation is essential to address the worldwide problems created by AI-generated misinformation and to ensure that AI supports contextual truth and accuracy instead of becoming a tool for deceit and mass manipulation.
The AI race is more than just competition between the United States and China because it represents a worldwide issue. By working together, democracies can leverage the transformative potential of AI while minimizing its risks and ensuring that it is used as a force for good in the world.
Dave Schroeder, MS, MA, CISSP, is a national security research strategist at the University of Wisconsin–Madison and research director for the Wisconsin Security Research Consortium. He also serves as a cyber warfare officer in the Wisconsin Army National Guard and previously served as a cryptologic warfare officer in the U.S. Navy. He can be reached at https://daveschroeder.org.
The views expressed here are the author’s alone and do not represent the views of any organization with which he has a current or prior affiliation.
