Competition, Not Control, is Key to Winning the Global AI Race
The United States has long been recognized as a front-runner in the arena of artificial intelligence (AI), housing much of the essential infrastructure that fuels AI development globally. However, this leadership position is under threat as China makes significant strides in state-of-the-art AI systems, challenging the U.S. dominance despite formidable U.S. efforts to curb China’s technological ascent through a series of export controls.
China’s rapid advancements in AI underscore a broader reality: the global AI competition is increasingly seen through the prism of U.S.-China rivalry. This competition is not just about who has the most advanced AI systems but also about who shapes the future direction of AI technology and its impacts on global geopolitics.
The U.S. response has involved a series of policy measures aimed at restricting Chinese access to critical components for AI advancement, including cutting-edge semiconductors and computational resources. Despite the intentions behind these export controls, their efficacy is debatable. Soon after the U.S. reinforced its export controls, China accelerated its efforts towards semiconductor self-sufficiency and firms found creative ways to circumvent these restrictions, illustrating the inherent challenges in controlling the flow of technology.
Looking ahead, the Biden administration is contemplating further export controls on AI at the software level. These proposed restrictions aim to limit China’s access to sophisticated “closed models” of AI, which lack publicly available source code. Yet, software’s intangible nature makes such controls particularly challenging to enforce, as software can easily cross borders undetected.
This shift towards increasingly stringent controls on AI software and related technologies raises significant questions about the long-term efficacy and strategic wisdom of these policies. As the AI landscape evolves, with open-source solutions diminishing the performance gap between open and closed-source models, the utility of targeting AI software with export controls is diminishing.
The pursuit of overly restrictive export controls may, in the end, be counterproductive for the United States. Potential consequences include eroding America’s competitive edge, decoupling from critical technology developments, and creating uncertainty within the domestic tech industry.
To maintain its global leadership in AI, the U.S. should pivot towards fostering competition and innovation. Supporting the widespread adoption and diffusion of American AI technologies both domestically and internationally can offer a more effective counter to the rise of digital authoritarianism than a control-centric strategy.
Key steps for U.S. policymakers include adopting a balanced approach to AI export controls, focusing on protecting critical applications without stifling innovation. Engaging international partners to set global standards for AI technology will also be vital, ensuring that American AI systems are adoptable worldwide. Moreover, recognizing the global nature of AI users and tailoring American AI technologies to meet diverse international needs can boost the U.S.’s global standing in AI.
Without a strategic shift towards embracing competition over control, the United States risks forfeiting its leadership position in AI to more assertive adversaries. By championing competition and innovation in AI, the U.S. can secure its technological supremacy and unlock the full potential of AI for itself and the world.
