Global Tech Policy Tracker — 2026-03-22

New Legislation & Proposals

United States: National AI Legislative Framework

• What it does: The White House published a four-page "National Policy Framework for Artificial Intelligence" on March 20, laying out the administration's vision for how Congress should legislate AI. The framework urges a federal-first, light-touch approach focused on preserving innovation and U.S. competitiveness against China.
• Who it affects: AI developers, tech companies operating across multiple U.S. states, and state legislatures currently pursuing their own AI bills.
• Timeline: Released March 20, 2026. Directed at Congress as a legislative wishlist; no binding legal force on its own.
• Key provisions:
  • Preempt state AI laws viewed as burdensome or conflicting with the federal approach, ending the current patchwork of state regulations.
  • Protect children online from AI-related harms, one of the few substantive regulatory priorities explicitly named.
  • Shield communities from high energy costs associated with AI infrastructure buildout.

1 sources

politico.com

politico.com

European Union: EU AI Act Simplification & Delays

• What it does: The European Parliament voted this week to support postponement of certain AI Act rules and backed proposals to simplify the regulation's requirements. MEPs also voted to propose bans on AI "nudifier" systems — tools that generate non-consensual intimate imagery — and to clarify application dates for high-risk system requirements.
• Who it affects: Companies deploying high-risk AI systems in the EU; developers of generative AI tools; any business subject to the AI Act's compliance obligations.
• Timeline: MEPs voted on March 18, 2026 (4 days ago). The measure now moves toward trilogue negotiations between the Parliament, Council, and Commission.
• Key provisions:
  • Postponement of certain rules, offering companies additional compliance runway.
  • Explicit ban on AI "nudifier" systems added to the text.
  • Clearer application dates established for high-risk AI system requirements.

The EU Council had also agreed just the week prior (March 13) on a position to streamline AI rules as part of the broader simplification push.

1 sources

europarl.europa.eu

europarl.europa.eu

Regulatory Actions

EU Digital Officials vs. Regulatory Complexity

• Action: EU officials spoke at a European Data Protection Board (EDPB) workshop this week on "regulatory interplay," discussing how digital simplification goals intersect with data protection and AI rules. Officials signaled that simplification does not mean abandoning convergence between GDPR, the AI Act, and other digital regulations.
• Significance: The workshop underscores growing tension inside the EU institutions between member states and industry groups pushing for deregulation and civil society advocates warning that core protections must not be sacrificed. With the AI Act omnibus heading to trilogue, the outcome will determine enforcement teeth for AI rules across 27 countries.

Industry Response

The White House framework landed exactly where major AI industry players wanted it. As CNBC reported, tech industry leaders have long opposed state-level regulatory efforts, arguing that a "patchwork" of laws would hobble innovation and hand China a competitive edge — language that appears almost verbatim in the administration's official framing.

On the EU side, critics are pushing back hard against the simplification drive. An analysis by TechPolicy.Press warned that "as the omnibus proposal moves toward trilogue negotiations, decision-makers must not lose sight of the AI Act's goals and effectiveness," arguing that watering down protections at this stage would undermine the regulation's fundamental purpose.

The Fortune framing of the U.S. framework — emphasizing "protecting children, preserving free speech, and cementing U.S. dominance in the global AI race" — reflects how the administration is packaging a broadly deregulatory agenda in consumer-friendly language likely to build bipartisan Congressional support.

What This Means

• State AI laws face an existential threat in the U.S. The White House framework's call to preempt state laws is a direct shot at efforts underway in California, Texas, Colorado, and other states. If Congress acts on this blueprint, companies that have built compliance programs around state-level requirements may need to pivot entirely to a federal standard — potentially a less demanding one.

• The EU AI Act is softening, but not disappearing. MEPs' vote to delay and simplify does not kill the regulation; it shifts timelines and reduces some compliance burdens. The ban on AI nudifier systems shows that targeted prohibitions remain politically viable even in a deregulatory moment. Companies should not interpret the delay as a reprieve from eventual compliance obligations.

• A global "race to the bottom" risk is emerging. With the U.S. explicitly framing light-touch regulation as a competitive strategy against China, and the EU easing its own rules in response to competitiveness concerns, there is a growing risk that substantive AI safeguards get traded away in a geopolitical competition narrative.

• AI startups face a bifurcated compliance reality. Operating in both the U.S. and EU now means navigating two increasingly divergent philosophies: Washington wants minimal federal rules and no state interference; Brussels is still building a comprehensive risk-based regime, even if its timelines slip. Dual-market companies will need to maintain EU-grade documentation and risk management practices regardless of what U.S. law ultimately requires.

Region Scorecard