Decoupling America's Artificial Intelligence Capabilities from China Act of 2025
- Sponsor
- Sen. Hawley, Josh [R-MO]
- Committees
- Judiciary Committee (primary)
- Last action
- Jan 29, 2025
Bottom line
The bill aims to severely restrict the flow of artificial intelligence technology, intellectual property, and expertise between the United States and China, particularly concerning entities linked to the Chinese government or military, to safeguard U.S. national security and technological leadership.
What it actually does
This bill establishes broad prohibitions on United States persons regarding artificial intelligence (AI) capabilities within the People's Republic of China. It bans the import of AI technology developed in China, the export of AI technology to China, and U.S. persons' engagement in AI research and development with Chinese entities of concern. Furthermore, it prohibits U.S. persons from holding financial interests in or providing financing to Chinese AI entities that contribute to China's military-civil fusion strategy, surveillance, or human rights abuses.
Proponents argue
Proponents argue this bill is crucial for national security, preventing China from leveraging U.S. AI advancements for military modernization, surveillance, and human rights abuses. They contend it protects U.S. technological leadership and intellectual property by strategically decoupling critical AI capabilities from a strategic competitor, thereby preserving American values and interests.
Opponents contend
Opponents might argue that such broad prohibitions could stifle legitimate scientific collaboration, harm U.S. companies operating in China, and potentially isolate the U.S. from global AI development, leading to a less innovative and competitive U.S. AI sector in the long run. They might also raise concerns about the practical enforceability and potential for unintended economic consequences, including retaliatory measures from China.
The bill is relatively short but dense with legal definitions and prohibitions, requiring careful reading to grasp its full scope and implications.
Section 1
Short Title
This section establishes the official short title of the Act as the 'Decoupling America's Artificial Intelligence Capabilities from China Act of 2025.' This name clearly indicates the bill's primary objective to separate U.S. AI capabilities from China.
Supporters argue
Supporters would argue that a clear and direct short title helps communicate the bill's purpose to the public and stakeholders, emphasizing its strategic importance.
Critics contend
No specific opposition is typically raised against short titles, as they are primarily for identification.
Section 2 and new Section 2741 of Title 18
Key Definitions for AI, Technology, and Chinese Entities
This section defines critical terms such as 'artificial intelligence,' 'generative artificial intelligence,' 'artificial intelligence or generative artificial intelligence technology or intellectual property,' 'Chinese entity of concern,' 'United States person,' 'military-civil fusion strategy,' and 'technology.' These definitions are foundational, determining the scope of all subsequent prohibitions. For instance, 'Chinese entity of concern' is broadly defined to include Chinese educational institutions, research labs, corporations headquartered in China, and any entity of the Government of the People's Republic of China, the Chinese Communist Party, or the People's Liberation Army.
Section 3
Ban on Import/Export of AI Technology and Intellectual Property with China
This section prohibits, 180 days after enactment, the importation into the U.S. of artificial intelligence or generative artificial intelligence technology or intellectual property developed or produced in the People's Republic of China. It also prohibits the export, reexport, or in-country transfer of such technology or intellectual property to or within the People's Republic of China. Violations carry criminal and civil penalties similar to those under the Export Control Reform Act of 2018.
Section 4, adding new 18 U.S.C. 2742
Ban on U.S. Persons' AI Research and Development with Chinese Entities
This section prohibits U.S. persons from intentionally conducting, attempting, conspiring, or aiding in AI research or development within China, for or on behalf of a 'Chinese entity of concern,' or in collaboration with such an entity or a Chinese national working for one. It also bans transferring AI research information under similar circumstances. Violations lead to substantial fines (up to $100 million for entities, $1 million for individuals), forfeiture of federal benefits, and ineligibility for federal financial assistance for 5 years. For individuals, it can also lead to immigration consequences.
Section 5
Ban on U.S. Financial Interests in Certain Chinese AI Entities
One year after enactment, this section prohibits U.S. persons from knowingly holding or managing an interest in, or lending money or extending credit to, a 'Chinese entity of concern' that conducts AI research/development or produces AI-incorporating goods, AND assists China's military-civil fusion strategy, surveillance capabilities, or is implicated in human rights abuses. The President is granted authority under the International Emergency Economic Powers Act (IEEPA) to implement and enforce this, with associated criminal and civil penalties.
The definitions of 'artificial intelligence' and 'generative artificial intelligence' are exceptionally broad, encompassing virtually any automated system, software, or process that uses computation to determine outcomes, make decisions, or interact with humans. This includes systems that 'approximate a cognitive task' or are 'intended to or capable of interacting with humans, detecting human emotions, determining associations with social categories based on biometric data, or generating or manipulating image, audio, video, or related material.'
Section 4, adding 18 U.S.C. 2741(1) and (8)
Why it matters:This breadth is likely intentional to ensure the prohibitions are future-proof and cover a wide array of potential AI applications, preventing loopholes. It's standard legislative practice to define terms broadly to achieve comprehensive coverage.
Case for: Proponents would argue that broad definitions are necessary to prevent circumvention and ensure the bill remains effective as AI technology evolves rapidly. Narrow definitions could quickly become obsolete.
Case against: Opponents might argue that such expansive definitions create legal uncertainty, potentially chilling legitimate research and commercial activities that are not directly related to national security concerns, due to fear of inadvertently violating the law.
Estimated impact: Significantly expands the scope of technologies and activities subject to the bill's prohibitions, impacting a vast segment of the technology sector and academic research.
The definition of 'United States person' is expanded beyond typical interpretations to include not only corporations and institutions organized under U.S. law, but also 'a person that is... a protected individual (as defined in section 274B(a)(3) of the Immigration and Nationality Act... wherever located or employed; and an alien who has been admitted to the United States as a lawful permanent resident.' It also includes 'a person or corporation controlled by an individual or entity described in this paragraph.'
Section 4, adding 18 U.S.C. 2741(14)
Why it matters:This broad definition is likely intended to ensure that individuals with strong ties to the U.S., regardless of their current physical location or citizenship status (if a lawful permanent resident), are subject to the prohibitions, preventing them from acting as intermediaries. This is a common tactic in national security legislation to prevent circumvention.
Case for: Proponents would argue that this comprehensive definition is crucial to prevent individuals with access to U.S. technology and knowledge from transferring it to Chinese entities, regardless of their current employment or location, thereby closing potential loopholes.
Case against: Critics might argue that this definition could create significant extraterritorial enforcement challenges and potentially lead to discrimination or undue scrutiny of lawful permanent residents or individuals with dual affiliations, even if their activities are not directly harmful to U.S. security.
Estimated impact: Extends the bill's prohibitions to a much wider array of individuals and entities globally, significantly increasing the compliance burden and potential legal exposure for individuals with U.S. ties.
This subsection amends the Immigration and Nationality Act to classify an offense described in new 18 U.S.C. 2742 (AI research and development on behalf of China) as an 'aggravated felony.'
Section 4(b), amending 8 U.S.C. 1101(a)(43)
Why it matters:Amending existing code sections is standard legislative practice. However, the placement here might make it less immediately apparent to those focused solely on the new criminal code provisions. It underscores the severity of the offense from an immigration perspective.
Case for: Proponents would argue that classifying such offenses as aggravated felonies is a necessary deterrent and ensures that individuals who compromise national security through AI collaboration with China face severe immigration consequences, including deportation.
Case against: Critics might argue that this adds another layer of punitive measures that could disproportionately affect non-citizen researchers or individuals with complex immigration statuses, potentially leading to harsh and irreversible consequences for actions that might be ambiguously defined or inadvertently committed.
Estimated impact: Creates a direct and severe immigration consequence (e.g., deportation, inadmissibility) for non-citizens convicted of violating the AI research prohibitions, significantly increasing the stakes for affected individuals.