US-China AI race, Innovation vs National Security debate intensifies

The United States faces intense competition with China in the field of artificial intelligence (AI) and former President Donald Trump’s technology allies have become divided over how the country should approach the issue.

According to a report by the U.S. news website Axios, supporters of AI in the United States have split into two camps. One group argues that strict controls on advanced AI are necessary to protect national security while the other believes such restrictions could cause American companies to fall behind China in the global AI race.

The debate has become increasingly public and could significantly influence how a future Trump administration regulates the world’s most advanced AI technologies.

Former Trump AI and cryptocurrency adviser David Sacks warned that limiting access to America’s most advanced AI models could undermine the strategy President Trump announced a year ago.

In a post on X, Sacks wrote: “A year ago, President Trump declared that America was in a global AI race and that the way to win it was by supporting innovation. President Trump was absolutely right. If we abandon that strategy, it will be dangerous for us.”

The comments came after reports that the White House had asked OpenAI to delay the broad release of its latest AI model, GPT-5.6 which is now expected to be rolled out in phases. Earlier, similar government guidance reportedly led Anthropic to suspend public access to its Fable 5 and Mythos 5 AI models.

According to a letter from Commerce Secretary Howard Lutnick, negotiations between the U.S. government and Anthropic have made significant progress. As a result, Mythos has become available again on a limited basis while Fable 5 is also expected to return soon.

Kevin Bankston, AI governance adviser at the Center for Democracy and Technology criticized the restrictions saying: “This is how you destroy the American AI market.”

Aaron Levie, CEO of Box said the AI industry has advanced rapidly because competing research labs have continuously pushed each other to build better models and that competitive pressure has been a major driver of innovation.

The report also noted that U.S. AI laboratories may face government-imposed limits on development speed while Chinese competitors are not subject to similar restrictions.

According to two separate security assessments, Chinese AI systems in the field of cybersecurity have already reached a level comparable to America’s leading AI models. The use of open-source Chinese AI models has also grown significantly in recent weeks largely because of their lower operating costs.

Platforms such as OpenRouter now rank several Chinese models among the most widely used AI systems. Investors and technology executives say they want clear, consistent regulations rather than case-by-case government decisions regarding access to advanced AI models.

Meanwhile, Anthropic continues to advocate for stronger safety measures as AI models become increasingly powerful arguing that advanced capabilities should be accompanied by robust safeguards.