The competition between the United States and China regarding artificial intelligence (AI) is entering a new phase and is increasingly focused on influencing the global AI ecosystem. Although national security is an important consideration in the U.S.-China AI rivalry, it represents only one dimension of a broader and nuanced competition. Both countries have released AI action plans that outline their respective strategies for supporting the development and dissemination of AI software and hardware. Notably, both plans call to develop open AI models, which, for the United States, represents a shift from closed or proprietary models to open-source models and environments. Both countries call for open AI in differing ways and levels of detail, but it is clear that open models are emerging as a critical aspect of each country wielding soft power.
 
In this paper, the authors aim to inform the United States’ open-source AI strategy by examining the diverging frameworks of the United States’ and China’s AI action plans. They evaluate how AI developers in both countries are approaching open models to gain influence and offer recommendations for U.S. policymakers to ensure that the United States maintains its technological leadership.
 
The authors argue that a strategy that emphasizes stronger support for an open model ecosystem in the United States, a recalibration of export controls, and the creation of incentives for firms to adopt permissive licensing practices is needed to counter the influence that China hopes to gain through diffusing its technology and tech infrastructure.

Read the full report at Rand.org.