Security Guarantees in an America First AI Trade Policy

As artificial intelligence rapidly emerges as history's most economically transformative technology, the United States stands at a critical juncture. Within this current Administration, we are likely to witness the dawn of artificial general intelligence—AI systems capable of performing virtually any economically valuable task. This technological leap could fundamentally reshape societies, creating a world where specialized AI chips power many essential economic activities.

The Administration possesses a rare strategic advantage: control over who accesses U.S.-manufactured AI semiconductors represents one of its most powerful levers in international trade discussions. But what precisely constitutes an America First AI Trade Policy? The Administration faces the difficult task of simultaneously advancing American technological and economic leadership while safeguarding national security interests. This challenge can be broken down into four core components.

Four Essential Properties of an America First AI Trade Policy

From the Administrations perspective, an effective America First approach to AI trade policy should accomplish four critical objectives:

1. Enable bilateral dealmaking leveraging America's AI chip advantage

2. Ensure global competitiveness of U.S. semiconductor manufacturers

3. Maintain the hardware edge for U.S.-based AI companies

4. Prevent adversarial use of U.S. technology in weapons development and other security threats

These properties represent the foundation of a trade policy that is supposed to put American interests first—but balancing them requires sophisticated policy mechanisms. Let's examine each in detail.

1. Enabling Bilateral Dealmaking Through AI Chip Access

AI is becoming the most economically powerful technology in history. Within this administration, we'll likely see artificial general intelligence emerge. This could lead to a world where AI systems running on specialized chips handle nearly all important economic work. The administration has a unique opportunity: controlling who gets access to U.S.-made AI chips can become one of its strongest tools in global trade negotiations. The increasing divergence in AI ecosystems highlights the importance of aligning with partners who share our commitment to democratic values and responsible AI development. This strategic alignment is crucial for ensuring the long-term security and integrity of global AI supply chains.

This advantage provides a basis for strategic cooperation, allowing the United States to engage in bilateral discussions that foster mutual economic growth and innovation across diverse sectors—from manufacturing to agriculture to services. By collaboratively managing access to these critical components, the administration can facilitate tailored agreements that benefit both American economic interests and the economic development of its partners.

2. Ensuring U.S. Semiconductor Global Competitiveness

Selling U.S.-developed and manufactured AI chips globally is essential for addressing trade imbalances and maintaining American dominance in semiconductor technology. The Trump administration's AI datacenter deals with the UAE and Saudi Arabia, worth billions of dollars, demonstrate how chip exports can significantly improve America's trade position. Furthermore, an overly restrictive approach to chip exports could inadvertently stimulate the acceleration of domestic semiconductor industries in competitor nations, potentially undermining long-term U.S. leadership.

American semiconductor giants like NVIDIA and AMD must maintain their global market position to ensure continued leadership in chip design and manufacturing. Restricting their ability to sell internationally would not only harm these companies financially but could ultimately diminish American innovation capabilities as competitors grow stronger. An America First policy recognizes that these companies' global success directly benefits the American economy and technological ecosystem.

3. Maintaining the Hardware Edge for U.S. AI Companies

However, to ensure the sustained leadership of American AI companies, it is crucial to manage the scale of access to advanced AI chips by foreign entities. This approach helps preserve the substantial advantages these U.S. firms currently hold in AI capabilities, revenue generation, and capital raised, which are largely due to their preferential access to cutting-edge AI hardware.

This hardware edge has been acknowledged not only by the CEOs of leading American AI companies but also by their international counterparts. The founder of Chinese AI company DeepSeek has publicly recognized that limited access to advanced computing resources has significantly hampered their ability to compete with U.S. firms. Maintaining this strategic advantage ensures that the economic benefits of AI development primarily flow to American companies, workers, and the broader U.S. economy.

4. Preventing Adversarial Misuse of U.S. Technology

Capable AI systems are recognized as a potential existential threat by the CEOs of all major AI labs, Nobel laureates, and the godfathers of modern AI. For this reason, it is critical that powerful AI capabilities don't fall into the hands of terrorists, dictatorial regimes, and U.S. adversaries. Scientists have warned that current AI models can help engineer dangerous pandemics, highlighting the concrete security risks involved.

While AI offers enormous economic and social benefits, advanced models and large computing clusters could enable adversaries and malicious actors to enhance military and intelligence applications, lower barriers to developing weapons of mass destruction, support sophisticated cyber operations, and assist in human rights violations such as mass surveillance. The United States must leverage its regulatory tools and authorities to mitigate these risks and protect national security.

Navigating Competing Priorities Through Technology

A fundamental tension exists between the first two properties (enabling bilateral dealmaking and ensuring semiconductor competitiveness) and the latter two (maintaining a U.S. AI hardware edge and preventing misuse). If universal access to U.S. AI semiconductors is provided, it risks eliminating the advantage enjoyed by American AI companies and lose control over potential misuse by adversaries.

The solution lies in implementing a proportional access framework: granting access to U.S. AI semiconductors based on the verifiable security guarantees that importers can provide. This approach allows the administration to finely calibrate its trade strategy, rewarding trusted partners while limiting access for less reliable actors.

For this proportional access framework to succeed, robust security guarantees must be implemented to ensure that AI chips remain with their intended recipients. Without proper safeguards, countries could potentially request more chips than needed and resell them to third parties or allow unauthorized remote access to these valuable computing resources. Such leakage would undermine America's bilateral negotiating power for what is rapidly becoming the world's most crucial technological resource.

1. Access Proportional to Security Guarantees

The current regulatory framework for critical technologies, notably AI, employs tiered control levels based on destination, end-user, and item sensitivity. While this differentiated approach is appropriate, the allocation of access (e.g., the number of imported chips) should be explicitly tied to the strength of verifiable, hardware-based security guarantees provided by the recipient. At a minimum, significant compute imports must necessitate verifiable proof of location within an authorized facility.

To safeguard the integrity of global AI supply chains, it is paramount that all trusted partners collaboratively uphold rigorous location verification requirements. This universal standard is not solely about preventing diversion, as evidenced by past instances of U.S. microchips reaching unauthorized destinations via intermediaries; it's about fostering a shared foundation of trust and security to protect our common strategic interests. Achieving these hardware-based guarantees can be significantly advanced through technologies such as confidential computing, trusted execution environments (TEEs), and zero-knowledge architectures.

This framework translates into crucial questions across the AI lifecycle, that could include:

• AI Inputs: "Where are the AI chips and other critical inputs located, and are global export controls being rigorously followed?"

• AI Training: "Is a trained model undergoing and passing specific safety evaluations before deployment, with auditable and tamper-resistant records?"

• AI Deployment: "Is the data of AI outputs and prompt logs being end-to-end encrypted, and in what jurisdiction is it stored, with assurances against unauthorized access?"

2. Development of Technical Standards

Collaboration between the Bureau of Industry and Security (BIS), the National Institute of Standards and Technology (NIST), and U.S. industry is needed to develop comprehensive technical standards and reference architectures for crucial security protections. Standardizing requirements for tiered access will clarify expectations and accelerate the development of compliant solutions.

These technical standards should prioritize performance-based, privacy-preserving location verification while also incorporating other vital security measures, such as comprehensive data center security controls to prevent unauthorized users or hackers from accessing advanced AI chips.

3. Independent Third-Party Auditing

Compliance with security standards must be verified through mandatory audits conducted by independent third parties. These auditors must be distinct from both U.S. exporters and foreign importers/end-users to ensure impartiality, similar to auditing practices in the financial sector. This independent verification is essential for the credibility and effectiveness of security requirements.

4. Funding and Competitiveness

The costs associated with implementing and auditing these crucial security guarantees will primarily be a responsibility of the foreign importers, reflecting their benefit from accessing advanced U.S. technology. Furthermore, the development of innovative hardware and software solutions required for compliance should foster growth among U.S. technology firms, particularly startups and established technology companies. This approach supports U.S. technological leadership in the critical field of secure data center infrastructure and AI security.

Conclusion

By implementing a system of proportional access based on verifiable security guarantees, the administration can balance competing priorities: enabling beneficial bilateral trade, maintaining U.S. semiconductor industry competitiveness, preserving the hardware edge for American AI companies, and preventing adversarial misuse of powerful technologies.

This balanced approach ensures that America's technological leadership translates into economic and strategic advantages while safeguarding national security. As AI continues its rapid development toward artificial general intelligence, the policies established today will determine whether the United States maintains its position at the forefront of this transformative technology—with all the economic and security benefits that leadership entails.

By insisting on robust security guarantees from trading partners, the administration can leverage America's AI chip advantage to secure favorable trade terms while preventing the proliferation of dual-use technologies to potential adversaries. This strategy represents the essence of an America First approach to AI trade policy: maximizing benefits to the American economy while minimizing risks to national security.