No Waiting for Washington: Maryland's Moore Unveils a State-Level AI Framework Built on Three Principles

While Congress continues to debate federal AI legislation without passing any, the states keep moving. On September 22, 2026, Maryland Governor Wes Moore stood in Annapolis and published something between a manifesto and a regulatory roadmap: a new AI framework organized around three principles — “protecting our people,” “centering our workers,” and “keeping our kids safe” — that his administration says will guide “every action” the state takes on artificial intelligence.

The timing is not subtle. “We can’t afford to wait while Washington sits on their hands,” Moore said in the announcement. With no comprehensive federal AI statute on the books, Maryland is joining a fast-growing club of states — New York ordered frontier labs to register with the state just last week under its RAISE Act — betting that if guardrails are coming to American AI, they will arrive state by state first.

What the framework actually says

The document is less a single law than an agenda spanning legislation, executive guidance, and existing statutes being reinterpreted for the AI era. Under its first principle, protecting Marylanders, the most consequential commitments include:

• Frontier model regulation. Maryland’s AI Subcabinet will develop legislative recommendations for regulating frontier AI model companies, including safety frameworks and testing, worker protections, independent third-party audits and evaluations, whistleblower protections, public reporting, and 72-hour incident reporting — a requirement that echoes disclosure norms the EU AI Act imposed on providers of high-risk systems.
• Your likeness as property. The administration wants a statutory Right of Publicity. Under current law, someone whose face or voice is cloned must prove harm to get relief; the new right would establish ownership from the outset, flipping the burden at a moment when voice cloning and AI avatars have become trivially cheap.
• Civil rights guidance for the algorithmic era. The Maryland Commission on Civil Rights, together with financial, housing, and labor regulators, will clarify that existing discrimination, consumer protection, and lending laws already reach AI-driven decisions.
• Housing algorithms in the crosshairs. Two separate actions target housing: prohibiting algorithmic tenant-screening tools that discriminate on protected characteristics in properties financed by the state, and working with the legislature to ban landlords from using shared pricing platforms to coordinate rents on non-public competitor data — a direct swipe at the algorithmic rent-collusion allegations that have produced lawsuits in several states.
• At-risk gamblers. A commitment to bar gambling operators from algorithmically identifying and targeting users showing problem-gambling patterns, and to restrict late-night and post-loss wager solicitations.

The second principle, centering workers, is where the framework gets more distinctive. Rather than treating AI displacement as an inevitability to be cushioned, Maryland is formalizing worker voice: a standing Labor-Management Committee will give unions representing state workers an institutional seat in how AI is deployed in government, and the administration pledges to work with the legislature to regulate the practices workers fear most — social scoring, tracking used to undermine organizing, and systemic bias. It pairs that with reskilling investments, and a “convene on worker transition” bringing unions, businesses, and higher education together.

The third principle, child safety, is arguably the most immediately felt. Maryland will strengthen protections around chatbot use by young people, work with the Attorney General and other states to curb AI-driven addictive design on platforms used by under-16s, require schools to implement safety protocols for any AI tool a student interacts with, and adopt a “do no harm” policy under which classroom AI tools must have demonstrated learning outcomes, supplement rather than replace educators, and — notably — come with family notification before use. The state education department will also issue model K-12 lesson plans on what AI is and is not.

The foundation underneath it

None of this arrives in a vacuum. The framework explicitly builds on Moore’s 2024 executive order on AI, the state’s Artificial Intelligence Governance Act of 2024, and a string of enacted legislation: SB 141 safeguarding elections from deepfakes, SB 360 giving victims of AI-generated intimate images legal recourse, SB 8 criminalizing AI impersonation, and HB 820 ensuring final insurance coverage decisions are made by qualified doctors rather than algorithms. Maryland already operates under an Age-Appropriate Design Code Act and an Artificial Intelligence Ready Schools Act.

The administration is also leaning on Moore’s role as co-chair of the National Governors Association’s Council of Governors’ Cyber Subcommittee, tying AI risk to critical-infrastructure defense — power, water, transportation, healthcare, financial, agricultural, chemical, and communications systems — and pressing for stronger state-federal coordination. On the opportunity side, Maryland this year joined the RAISE US national initiative to prepare workers and employers for an AI-driven economy, layered on a $4 million Department of Labor investment in hands-on training for the state’s “lighthouse industries” and local cybersecurity programs.

Why it matters beyond Maryland

The significance is structural, not just local. First, the frontier-model regulation item puts a mid-Atlantic state on the same path New York opened with the RAISE Act — direct state oversight of frontier AI companies, including audit and incident-reporting duties that labs have so far only faced in Europe. If even a handful of large states adopt similar regimes, de facto national standards emerge without Congress lifting a finger, and compliance baselines (like 72-hour incident reporting) become table stakes for any lab operating in the U.S. market.

Second, the Right of Publicity proposal addresses one of the most lopsided fights in AI: individuals whose likenesses are harvested for training and cloning have had to litigate harm after the fact. Establishing likeness as a property right from the outset would give them standing before damage is done — a model other states watching the voice-cloning and avatar economy will study closely.

Third, the worker provisions test a hypothesis: that AI transitions can be governed with labor at the table rather than managed as layoffs with severance. Moore was candid in recent interviews that Maryland workers “have reason to be concerned” about AI; this framework is his answer to what the state will actually do about it — bargaining structures, regulated uses, and training pipelines rather than a specific displacement threshold, a choice critics on both sides will parse in the months ahead.

The open questions are the usual ones for principles documents: which pieces survive the legislature, how the AI Subcabinet defines “frontier,” and whether enforcement resources match ambition. But as a statement of direction, Maryland has just published one of the most comprehensive state AI frameworks yet — workers, children, housing, gambling, elections, and frontier labs all in one document, with the explicit argument that in 2026, states can no longer wait for Washington.