Three big developments this week. The Commerce Department published its report on state AI laws. The EU Council voted to delay key AI Act deadlines. And Colorado’s landmark AI law got pushed back again. Here’s what’s happening.
Commerce Department Report: States in the Crosshairs
The Commerce Department released its state AI laws evaluation on March 11, fulfilling a December 2025 executive order. The report identifies categories of state regulation that the federal government considers problematic:
What’s flagged:
- Laws requiring AI systems to “alter or suppress truthful outputs”
- Transparency requirements that may reveal proprietary information
- Algorithmic discrimination rules in employment, housing, and credit
- Generative AI training data disclosure mandates
- Restrictions on AI-generated political content and deepfakes
The report doesn’t name specific states, but policy discussions have focused on Colorado, California, and New York. Republican representatives from Colorado and New York wrote to Commerce Secretary Howard Lutnick in February urging the department to include their states’ laws in the review.
What happens next:
The Department of Justice established an AI Litigation Task Force in January to challenge state AI laws in federal court. The Commerce Department can also condition $42 billion in broadband infrastructure funding (the BEAD program) on states repealing AI regulations deemed “onerous.”
The legal basis for preemption remains untested. Executive orders don’t automatically override state law, and courts may be skeptical of claims that state AI regulations violate interstate commerce. But the threat of losing federal funding gives states strong incentive to reconsider.
EU Council Votes to Delay AI Act Deadlines
The EU Council adopted its negotiating position on AI Act amendments on March 13. The proposal is part of “Omnibus VII,” a broader package to simplify EU digital regulation.
Key changes:
- High-risk AI system rules delayed by up to 16 months
- New deadline contingent on standards and technical tools being ready
- Regulatory exemptions extended from SMEs to small mid-cap companies
The August 2026 enforcement deadline for high-risk AI systems was always ambitious. As of this month, only eight of 27 member states have designated their required enforcement authorities. The delay acknowledges that neither regulators nor industry are ready.
The European Parliament still needs to set its position before trialogue negotiations begin. But the direction is clear: the EU is pulling back from aggressive timelines to focus on practical implementation.
For companies operating in Europe, this means more time to prepare—but also continued uncertainty about final requirements.
Colorado: Pushed to June, Future Uncertain
Colorado’s AI Act, the first comprehensive state AI law in the US, got delayed again. Governor Jared Polis signed SB 25B-004 in August 2025, pushing implementation from February 1 to June 30, 2026.
The law requires “reasonable care” to protect consumers from algorithmic discrimination in high-risk AI systems—employment, housing, credit, healthcare, and similar domains. Covered entities must conduct impact assessments and provide disclosures.
The problem: Neither the Governor nor the bill’s original sponsor seem to want the law as written. A working group has been meeting weekly on a potential repeal-and-replace bill. Consensus language is expected sometime this month.
If Colorado significantly weakens its law—or repeals it entirely—the federal preemption threat becomes less relevant. But it also signals to other states that comprehensive AI regulation is politically difficult to sustain.
What Else Is Moving
State legislatures:
- Washington’s HB 2225 (chatbots) and HB 1170 (provenance) passed both chambers and await the Governor’s signature
- Arizona’s SB 1786 passed the Senate 17-13
- New York’s AI Training Data Transparency Act advanced to floor readings
California: The AI Transparency Act and Generative AI Training Data Transparency Act took effect January 1. These require disclosure of AI-generated content and public summaries of training datasets. The Commerce Department report may target these laws, but California has historically resisted federal pressure on tech regulation.
What This Means
The regulatory landscape is fragmenting in unusual ways.
Federal vs. State: The Trump administration is using executive authority and funding leverage to push back against state AI laws, but the legal limits of this approach haven’t been tested. Companies face genuine uncertainty about which rules will survive.
EU vs. Industry: Europe is acknowledging that aggressive AI Act deadlines don’t match implementation reality. The delay buys time but doesn’t resolve underlying questions about compliance costs and technical feasibility.
Pioneers vs. Skeptics: Colorado’s retreat suggests that even states committed to AI regulation are finding the details harder than the principles. The gap between passing an AI law and enforcing one is widening.
What You Can Do
If you operate in the US: Track the DOJ’s AI Litigation Task Force. The first federal challenges to state AI laws will set important precedents.
If you operate in the EU: The August 2026 deadline for high-risk AI systems is likely moving. Focus on the updated timeline once Parliament acts, but don’t assume unlimited delays.
If you’re subject to Colorado’s law: Watch for the working group’s consensus language this month. June 30 is the current deadline, but the substance of compliance may change significantly.
The only certainty in AI regulation right now is uncertainty. Build compliance processes that can adapt.