U.S. AI strategy upgrade: obvious intention to encircle China, domestic substitution imminent?

White House Releases Ensuring U.S. National Security and Economic Strength in the Age of Artificial Intelligence

Artificial intelligence is rapidly becoming a centerpiece of security and economic power. The United States must act decisively to lead this transformation by ensuring that U.S. technology underpins the use of AI globally and that adversaries cannot easily misuse advanced AI. In the wrong hands, powerful AI systems have the potential to exacerbate significant national security risks, including enabling the development of weapons of mass destruction, supporting robust offensive cyber operations, and contributing to human rights abuses such as mass surveillance. Today, worrisome states are actively using AI - including U.S.-made AI - to act in this way and seek to undermine U.S. leadership in AI.

In order to enhance U.S. national security and economic strength, it is critical that we do not offshore this critical technology and that the world's AI is kept on U.S. tracks. It is important to work with AI companies and foreign governments on critical safety and trust standards as they build their AI ecosystems.

In an effort to strengthen U.S. national security and economic power, the Biden-Harris Administration today issued an interim final rule on the proliferation of artificial intelligence. It simplifies licensing hurdles for large and small chip orders, strengthens U.S. leadership in AI, and clarifies to allies and partner nations how they can benefit from AI. It builds on previous chip controls by discouraging smuggling, closing other loopholes, and raising AI security standards.

Six key mechanisms in the rule promote the responsible diffusion of U.S. technology:

• No restrictions apply to chip sales to 18 key allies and partners . This flexibility allows jurisdictions with strong technology protection mechanisms and technology ecosystems aligned with U.S. national security and foreign policy interests to benefit from seamless large-scale procurement.
• Chip orders totaling approximately 1,700 advanced GPUs of computing power do not require a license and do not count toward the state chip cap. The vast majority of chip orders fall into this category, particularly those placed by universities, medical institutions and research organizations for apparently innocuous purposes. Streamlining the processing of these orders would be an improvement over the status quo and could quickly expedite the low-risk transportation of U.S. technology around the world.
• Entities that meet high security and trust standards and are headquartered in close allies and partners can obtain the highly trusted Universal Verified End User (UVEU) status. With this status, they can place up to 7% of their global AI computing power in countries around the world - potentially equivalent to hundreds of thousands of chips. This trusted status is granted globally and permanently, allowing responsible entities to scale quickly and flexibly and strengthen the global leadership of the United States and allies while keeping cutting-edge training at home.
• Entities that meet the same security requirements and are headquartered in any country that is not a country of concern can apply for Nationally Validated End-User (NVEU) status, which will allow them to purchase the equivalent of up to 320,000 advanced GPUs of computing power over the next two years. This provision allows trusted national entities to benefit from advanced U.S. technology to serve local, governmental, and regional customers while guarding against transfer risks.
• non-VEU entities located outside of close allies can still purchase significant computing power. This cap ensures that U.S. technology can be used to serve foreign governments, healthcare providers, and other local businesses.
• Intergovernmental arrangements to foster an international ecosystem of shared values on the development, deployment and use of artificial intelligence. Governments that sign on to these arrangements-which align these countries' export control, clean energy, and technology security efforts with those of the United States-can double their chip caps (up to 100,000 of today's advanced GPUs).

Even as it encourages the proliferation of U.S. technology, the rule takes significant steps against countries of concern by restricting their access to advanced AI systems and the computing power used to train them. These measures include:

• Continuing to ensure that advanced semiconductors sold abroad are not used by countries of concern to train advanced artificial intelligence systems. while still allowing for general-purpose applications ranging from telecommunications to banking.
• Restrictions on transferring model weights from advanced closed weight models to untrusted participants. The rule will never prohibit the release of model weights for open weight models.
• Setting Safety Standards to Protect the Weights of Advanced Closed Weight AI Models , allowing them to be securely stored and used globally while helping to prevent access by illegal adversaries.

The rule builds on prior regulations designed to protect U.S. national security, including the October 2022 and October 2023 chip controls. It was developed after extensive and relevant engagement with stakeholders, members of Congress from both parties, industry representatives, and foreign allies and partners over the past ten months.

Related reading:

Smart Spectrum: Statement on Entity Listing by the U.S. Department of Commerce