The US-China Interplay Of Open Source AI Adoption – OpEd
China and the United States currently lead global advancements in Artificial Intelligence, and the strategic decisions taken by both countries are interdependent in some areas. China’s adoption of open-source AI models is increasingly influencing policymakers and market leaders in the U.S. to take a similar path, but there is also a scenario where the U.S. largely sticks to the status quo and influences China.
China’s open source policy has been adopted by the government, leading AI companies and labs in China to release their models using licenses that allow users to download and build upon the models. The performance of these models even competes with ranked ones in the U.S.
Influenced by the Chinese open source AI policy and the success of the DeepSeek, many voices in the U.S. are emerging, calling on the U.S. government and companies to follow this path. For instance, ex-CEO of Google, Eric Schmidt, warned that it’s time to focus on open source AI to compete with China. Researchers and activists are advocating for the U.S. to regain its leadership through open source AI. NVIDIA, the largest chip producer, launched an open-source simulation tool for robotics. XAI announced that it will open source its Grok2.5 model, and most recently, the U.S. AI action plan encouraged the adoption of open source AI.
The global adoption of Chinese models and its dominance in the different markets will directly provide China with a great opportunity to set the standards for the AI industry, one of which is making open source the default, which will put a lot of pressure on AI in U.S. companies to open up their models.
However, AI companies in the United States may not genuinely embrace open-source principles and could instead adopt a marketing strategy of selectively sharing models that are not fully open. This practice, often referred to as *open washing*, allows companies to claim transparency and openness while maintaining significant control over access and deployment. Meta’s recent model releases exemplify this approach, illustrating how commercial interests can shape the boundaries of openness in the AI ecosystem.
One scenario that should not be overlooked is the possibility that the United States will continue relying on closed AI models and succeed in proving their effectiveness in protecting society from potential misuse of open systems. By reducing risks related to the release of models that malicious actors could exploit to cause harm, the U.S. could confirm the viability of the closed-model approach. This outcome might also lead China to reconsider the long-term feasibility of its open-source AI policy.
Like what you read?
Please consider supporting Eurasia Review. Thank you for your consideration!
