Protection Against Foreign Adversarial Artificial Intelligence Act of 2025
- Bill Number
- S. 1638
- Origin Chamber
- Senate
- Congress
- 119th Congress, Session 1
- Policy Area
- Science, Technology, Communications
- Status
- Introduced
- Latest Action
- 2025-05-07: Read twice and referred to the Committee on Commerce, Science, and Transportation.
- Last Updated
- 2025-08-01T11:03:20Z
AI-Generated Summary
Purpose
The legislation aims to safeguard U.S. national security by restricting the use of certain artificial intelligence (AI) applications linked to foreign adversaries and requiring a detailed assessment of broader threats from AI platforms associated with "countries of concern" (nations identified as potential security risks, such as those covered under existing U.S. defense laws, often including China, Russia, Iran, and North Korea).
Key Provisions
- Prohibition on Specific AI Use in Federal Contracts (Section 2): Federal contractors with active contracts are banned from using the DeepSeek application—or any similar tools developed by High Flyer or its affiliates—to support or fulfill government contracts. This includes any role in execution, assistance, or partial completion of the work.
- Waiver Option: The Secretary of Commerce, in consultation with the Secretary of Defense, can grant exceptions on a case-by-case basis if the AI is essential for national security goals or legitimate research purposes.
- Required Report on AI Threats (Section 3): Within one year of enactment, the Secretary of Commerce (consulting the Secretary of Defense) must submit a comprehensive report to key congressional committees (Armed Services, Commerce/Science/Transportation, and Energy/Commerce).
- Report Contents:
- Analysis of how governments in countries of concern might access or influence AI apps through censorship laws.
- Examination of AI's role in spreading state-sponsored propaganda.
- Assessment of how these countries evade U.S. export controls on critical hardware (like graphics processing units) to build advanced AI models.
- Detailed review of privacy and data security risks for U.S. users' data in foreign AI systems, including:
- Storage locations (e.g., on-site servers vs. cloud).
- Potential access by foreign governments or entities (e.g., the Chinese Communist Party).
- How U.S. data might train or improve these AI models.
- Risks of economic espionage (stealing trade secrets, intellectual property, or sensitive info for unfair advantages).
- Threats to U.S. government policy or program information.
- Any additional relevant details.
- Recommendations for new rules or laws to mitigate data security and privacy risks from AI tied to adversarial governments.
- Format: The report must be unclassified but can include a classified (restricted-access) section for sensitive information.
Significant Changes to Existing Law
- Introduces a targeted ban on a specific foreign AI tool (DeepSeek) in federal contracting, which builds on but expands existing restrictions on foreign tech in government work (e.g., under defense procurement rules).
- Mandates the first-of-its-kind congressional report specifically focused on national security threats from foreign AI platforms, including large language models (AI systems that generate human-like text) and generative AI (tools that create images, videos, or other content). This fills a gap in current laws, which address general export controls and data security but lack AI-specific threat assessments.
Potential Impacts
- On Government Agencies: Federal agencies (e.g., Commerce and Defense Departments) must enforce the DeepSeek ban, potentially increasing compliance costs for contractors and requiring new oversight. The report could lead to stricter AI procurement guidelines, affecting how agencies adopt technology.
- On Citizens: Enhances data privacy protections by highlighting risks of sharing information with foreign AI, potentially reducing exposure to espionage or propaganda. However, it may limit access to certain AI tools for research or personal use if broader restrictions follow.
- On International Relations: Could strain ties with countries of concern by signaling U.S. distrust of their AI tech, possibly prompting retaliatory measures on U.S. exports or collaborations. It reinforces U.S. efforts to curb technology transfers, aligning with ongoing trade tensions.
Main Stakeholders Affected
- Federal Contractors and Businesses: Directly impacted by the DeepSeek prohibition, requiring them to switch AI tools and verify compliance to avoid contract penalties.
- U.S. Government Entities: Commerce and Defense Departments lead implementation and reporting; congressional committees receive and act on findings, influencing future policy.
- AI Developers and Users: Companies like High Flyer (and similar foreign entities) face U.S. market barriers; American researchers, developers, and citizens using AI may see reduced options or heightened scrutiny on data handling.
- Countries of Concern and Their Governments: Targeted for AI affiliations, potentially facing export restrictions or diplomatic pushback.
Notable Legal, Constitutional, or Political Implications
- Legal: Strengthens national security frameworks by integrating AI into federal contracting rules, potentially setting precedents for regulating emerging tech under existing espionage and export control laws (e.g., those in Title 10 of the U.S. Code). Waivers provide flexibility to avoid overly rigid enforcement.
- Constitutional: Raises questions about balancing national security with free speech or innovation rights, though the focus on federal contracts (not private use) minimizes First Amendment concerns. Data privacy analyses could support future protections under laws like the Fourth Amendment (unreasonable searches).
- Political: Bipartisan sponsorship (by Senators Cassidy and Rosen) suggests broad support for countering foreign AI threats, but the report's recommendations could spark debates on overregulation vs. security needs, influencing U.S. tech policy in a global AI race.
This summary was generated by AI and may contain inaccuracies. Refer to the official source document for the authoritative text.
Sponsor
Cosponsors (3)
Sen. Rosen, Jacky [D-NV], Sen. Schatz, Brian [D-HI], Sen. Budd, Ted [R-NC]
Recent Actions
- 2025-05-07: Read twice and referred to the Committee on Commerce, Science, and Transportation.
- 2025-05-07: Introduced in Senate
Bill Versions
- Protection Against Foreign Adversarial Artificial Intelligence Act of 2025 — issued 2025-05-07 — PDF (5 pages)