The White House says voluntary partnerships, not strict mandates, are the right approach to regulating AI.
summary
- The White House released the National AI Policy Framework in March 2026, favoring voluntary technology agreements over mandatory federal rules.
- The framework directs Congress to preempt state AI laws deemed to impose undue burdens on innovation and industry.
- Democratic lawmakers introduced the GUARDRAILS Act to block federal preemptions and preserve state authority to censor artificial intelligence.
Trump administration Released Its national policy framework for AI regulation in March 2026 is built on voluntary industry agreements rather than top-down mandates.
The Framework signals a move away from prescriptive regulation towards an innovation-friendly approach, positioning itself as a clear alternative to EU AI law.
The framework cites the March 2026 Taxpayer Protection Pledge, a voluntary agreement signed by big tech companies not to raise household electricity bills, as a model for the partnership-first approach it favors over binding rules. The administration’s basic premise is that US leadership in AI depends on uniform national standards, not an increasing patchwork of state laws.
Federal preemption versus state authority
The framework sets out six goals: protecting children online, protecting against the harms of artificial intelligence, respecting intellectual property, preventing censorship of artificial intelligence, encouraging innovation, and developing an artificial intelligence-ready workforce.
It calls on Congress to adopt legislation that broadly preempts state AI laws seen as imposing undue burdens, while preserving state authority regarding consumer protection, child safety, and fraud. Critics argue that this approach could hollow out oversight of high-risk AI systems in the fields of health care, employment and housing.
Democrats I paid Return directly. Representative Beyer and his colleagues introduced the GUARDRAILS Act on March 20, 2026, which would rescind the Trump administration’s executive order on AI and prevent any federal moratorium on AI regulation at the state level. Senator Schatz is expected to introduce companion legislation in the Senate.
What changes and what doesn’t
The Framework does not itself create new legal obligations or direct agencies to take specific regulatory actions. State AI laws remain in effect unless Congress passes new legislation or the courts strike them down.
Colorado’s comprehensive AI law is scheduled to go into effect on June 30, 2026. California’s AI Transparency Act and Texas’ Responsible AI Governance Act have already gone into effect, each imposing disclosure and governance requirements on companies deploying AI in subsequent decisions.
The CFTC has it separately Spread out Artificial intelligence tools to fill regulatory oversight gaps as broader framework battle continues in Washington. The administration has not said whether it will directly challenge active state laws, leaving companies navigating two parallel and perhaps conflicting regulatory paths.
