Reporter: “DeepSeek, do you believe it is a national security threat?”
President Trump: “No, I think it’s happening, it’s a technology that’s happening, we’re gonna benefit, it will be a lot less expensive than people originally thought, I view that as a very good development, not a bad development.”
This forthright, pragmatic statement about DeepSeek – an innovative, high performing, open source Chinese reasoning AI model which came as no surprise to followers of the field – is the latest signal from Trump 2.0 about US AI policy. Other actions include:
- Rescinding President Biden’s Executive Order 14110. The EO tasked the Department of Commerce with applying the concept of marginal risk analysis to dual-use and open source foundational models, and imposed rules on companies developing models and computing clusters. It also directed the USG to proactively promote progress, innovation, and competition in domestic AI development to fulfill national security objectives, including attracting global technical talent.
- Issuing an America First Trade Policy Memorandum. It called upon the Secretaries of State and Commerce to “review the United States export control system and advise on modifications in light of developments involving strategic adversaries or geopolitical rivals as well as all other relevant national security and global considerations.”
- At a summit of CEOs and heads of state in Paris, Vice President Vance said “We believe that excessive regulation of the AI sector could kill a transformative industry,” and the U.S. refused to sign the final statement of the summit that said AI should be inclusive, open, ethical and safe.
Together, Trump’s interventions seem to represent a shift in the way the US approaches AI as an emerging technology, economic competition between AI firms, and its relation to national security. The key question is: Can Trump 2.0 curb the bipartisan tendency in Congress to prioritize national security over economic competition in AI?
The Legacy of National Security-Driven AI Policy
Under both the prior Trump and Biden administrations, U.S. policy increasingly restricted AI-related technologies, curtailing global competition in the name of national security. Key measures included:
- Export Control Reform Act of 2018, which laid the foundation for the creation of a global advanced semiconductor control regime, first established in 2022, and expanded in 2023 and 2024;
- Foreign Investment Risk Review Modernization Act of 2018, curtailing newly covered transactions and inbound capital flows;
- Creation of Team Telecom and novel application of Section 214 of the Communications Act of 1934, to remove foreign-owned (Chinese) service providers from the U.S. market;
- Biden Executive Order 14105, grounded in the International Emergency Economic Powers Act and National Emergencies Act, to restrict outbound capital flows, and;
- Protecting Americans from Foreign Adversary Controlled Applications Act which may yet ban or force an ownership change in a market-leading platform over concerns of data security.
Escalating Restrictions: The 2025 Decoupling Act
More recently, extreme economic nationalist Senator Hawley introduced the Decoupling America’s Artificial Intelligence Capabilities from China Act of 2025. Having dropped any pretense of risk analysis informing policy, this bill threatens open source software underlying AI systems and broader firm economic value, by
- Prohibiting the import and export of “AI technology” to or from China
- Restricting AI research and development collaborations with Chinese entities,
- Prohibiting investment and financing in certain Chinese AI-related entities.
Additionally, bipartisan leadership of the House Select Committee on the Strategic Competition Between the United States and the Chinese Communist Party wants the White House to double down on already questionably successful controls which are having the opposite effect of making Chinese companies more competitive. It urged National Security Advisor Mike Waltz “to review the potential national security benefits of placing additional export controls on semiconductor chips critical to the AI infrastructure of People’s Republic of China (PRC) company, DeepSeek.”
A Business-Friendly White House or More Tech Protectionism?
How will the allegedly business-friendly White House respond? Could Trump’s approach be substantively different, favoring innovation, competition, and economic interests over national security concerns? Or is it simply framing a different, transactional version of tech protectionism (e.g., controls concessions for market access)? Some Departments understand that the latest shock to the AI ecosystem presents an opportunity for American firms that will also help them. Global AI provider platforms and chip designers were generally enthusiastic about the opportunity surrounding DeepSeek’s innovation too. American-Chinese researcher collaboration continues to grow. It remains to be seen whether the Trump 2.0 administration will embrace more market-oriented AI policies and refine national security-related controls to be more narrowly focused, or if it will continue using controls and restrictions as a geopolitical tool.
