The Introduction of the No Adversarial AI Act
In a notable show of bipartisan support, U.S. lawmakers have rolled out the “No Adversarial AI Act.” This legislation aims to create a robust barrier between federal agencies and artificial intelligence technologies associated with foreign adversaries.
The initiative is led by Raja Krishnamoorthi (D-IL), the ranking member of the House Select Committee on the Strategic Competition Between the United States and the Chinese Communist Party, alongside committee Chairman John Moolenaar (R-MI). Their goal is to address national security concerns, particularly regarding artificial intelligence systems that might be influenced by hostile foreign powers, with a primary focus on the People’s Republic of China (PRC).
The measure has gained support from both houses of Congress. Representatives Ritchie Torres (D-NY) and Darin LaHood (R-IL) have joined as cosponsors in the House, while Senators Rick Scott (R-FL) and Gary Peters (D-MI) have introduced a corresponding bill in the Senate.
AI as a Security Concern
The key premise behind the “No Adversarial AI Act” is that while AI has transformative potential, it can also serve as a tool for espionage when exploited by malicious entities. Concerns have been raised about companies such as DeepSeek, which reportedly has connections to the Chinese Communist Party and explicitly states that U.S. user data is stored in China. This raises alarms about data privacy and security.
Prior investigations led by Krishnamoorthi and Moolenaar suggested prohibiting federal procurement of AI models originating from the PRC, particularly for deployment on government devices. This legislation directly responds to those recommendations.
Key Features of the Bill
The “No Adversarial AI Act” employs a multi-faceted strategy to protect federal systems, encompassing several crucial provisions:
- Establishing a Federal List of Adversarial AI: The act obliges the Federal Acquisition Security Council (FASC) to compile a list of AI technologies developed by foreign adversaries within 60 days of its passage. This list will be publicly published by the Director of the Office of Management and Budget (OMB) within 180 days and updated every 180 days thereafter.
- Prohibiting Federal Use of Listed AI: Once an AI technology is included on this list, government agencies will be barred from procuring or utilizing it. This restriction extends to companies with known affiliations to the Chinese Communist Party, such as DeepSeek.
- Limited Exceptions with Strict Oversight: While the general rule is to prohibit these technologies, exceptions can be made. Heads of executive agencies can approve usage for essential scientific research, counterintelligence activities, or to safeguard mission-critical functions. However, they must inform the OMB and relevant congressional committees.
- Regular Updates and Removal Process: The FASC will need to refresh the adversarial AI list at least every 180 days, along with a methodology for removing entries if they can verify that technologies are not developed by foreign adversaries.
- Empowering Agency Enforcement: The legislation empowers executive agencies to utilize their existing authorities to exclude and eliminate AI provided by entities on the list classified as foreign adversaries.
Defining Foreign Adversary AI
The act adopts a broad definition of “artificial intelligence,” consistent with existing U.S. law, and clarifies what constitutes a “foreign adversary” and “foreign adversary entity.” The term encompasses:
- A foreign adversary itself.
- A foreign person based in a foreign adversary country.
- An entity wherein foreign individuals own 20% or more shares.
- A person under the influence of any of the above.
The definition of “foreign adversary” aligns with established U.S. legal references, typically including nations such as China, Russia, Iran, and North Korea.
Addressing Digital Warfare
Lawmakers are pushing for swift approval of this legislation. Krishnamoorthi emphasized that AI controlled by foreign adversaries poses a serious risk to national security and data integrity. He called for a decisive protective measure to guard U.S. institutions and citizens from hostile foreign infiltration.
Moolenaar echoed these sentiments, asserting that we are entering a new Cold War characterized by the strategic significance of AI technologies. He criticized the Chinese Communist Party’s approach to AI, alleging that their methods involve theft and manipulation of sensitive data.
Both Senators Scott and Peters expressed concerns about the risks associated with government reliance on foreign-controlled platforms. Scott warned about the dangerous implications of such ties to China and highlighted the need for heightened vigilance over user data security.
Peters underlined the act’s role in protecting U.S. systems from AI technologies that could compromise national security. This legislation represents a pivotal step in securing the U.S. digital landscape against evolving geopolitical challenges.
