Congress AI Policy: Two Bills in Final 2025 Session

Congress convened its final session of 2025 on December 2 with healthcare topping the agenda—but two AI policy measures have a last chance for action before the session closes.

US AI regulation has lagged behind the EU and other jurisdictions. These bills represent incremental steps, not comprehensive frameworks.

The Current AI Policy Landscape

Unlike the EU's AI Act or China's AI regulations, the United States lacks comprehensive federal AI legislation. Current governance relies on:

• Executive Orders: Biden's October 2023 AI executive order established reporting requirements and safety standards for federal use
• Agency Guidance: FTC, EEOC, and other agencies applying existing laws to AI
• State Laws: Patchwork of state-level AI regulations
• Industry Self-Regulation: Voluntary commitments from AI companies

The bills under consideration address specific issues rather than establishing broad frameworks.

What's Being Debated

The two AI measures in the final session target narrow concerns:

AI in Government Operations Requirements for federal agencies deploying AI systems, including:

• Transparency about AI use in decision-making
• Risk assessments for high-impact applications
• Human oversight requirements
• Reporting obligations to Congress

AI Research and Development Provisions affecting:

• Federal AI research funding priorities
• Coordination between agencies on AI development
• International AI research collaboration

Neither bill approaches the scope of comprehensive AI regulation.

Why Comprehensive Legislation Stalled

Several factors prevented broad AI legislation in 2025:

Definitional Challenges: What counts as "AI"? Definitions that are too broad capture simple software; too narrow miss emerging capabilities.

Industry Opposition: Tech companies lobbied against regulations that could constrain innovation or create compliance costs.

Partisan Divides: Republicans generally favor light-touch approaches; Democrats often prefer stronger consumer protections.

Rapid Technology Change: Legislators struggled to write rules for technology evolving faster than legislative processes.

Jurisdictional Complexity: Multiple committees claim AI oversight—Commerce, Judiciary, Intelligence—complicating unified approaches.

Comparison to EU AI Act

The EU's AI Act, taking effect in stages through 2026, establishes:

• Risk-based classification system for AI applications
• Prohibited AI uses (social scoring, certain biometric systems)
• Requirements for high-risk AI systems
• Transparency obligations
• Penalties for non-compliance

US legislation under consideration is far narrower, addressing government AI use rather than private sector deployment.

State-Level Action

While federal legislation stalled, states moved forward:

California: Multiple AI bills covering algorithmic discrimination, deepfakes, and AI safety Colorado: AI algorithmic discrimination law New York City: Local Law 144 requiring bias audits for employment AI Illinois: Biometric information protection

This patchwork creates compliance complexity for companies operating nationally.

Industry Implications

For AI Companies: Continued federal regulatory uncertainty. Planning requires monitoring state laws and potential future federal action.

For Enterprises: Government contractors face new AI requirements if bills pass. Private sector deployment remains largely unregulated federally.

For International Operations: US companies operating in EU face AI Act compliance regardless of US legislative outcomes.

What Passes (Probably)

Observers expect the government-focused measures have better passage prospects than private sector regulation. Federal government as customer for AI creates bipartisan interest in appropriate guardrails.

R&D provisions typically attract support—both parties generally favor American AI competitiveness.

Looking to 2026

The new Congress in January 2026 will likely revisit AI policy. Several factors could accelerate action:

• Major AI incident creating public pressure for regulation
• International pressure as EU AI Act implementation proceeds
• Industry support for federal preemption of state laws
• Electoral dynamics making AI a voter issue

Comprehensive US AI regulation remains uncertain but increasingly likely as AI capabilities and deployment expand.

For Organizations

Government Contractors: Monitor final session outcomes; prepare for potential new requirements.

Private Sector: Expect continued state-level regulatory activity; federal action more likely in 2026.

All Organizations: Document AI systems and decisions now; regulation requiring retrospective information will be easier to satisfy.

The absence of federal AI regulation is itself a policy choice—permitting innovation while accepting associated risks. Whether that balance shifts depends on both political dynamics and AI's real-world impacts.