AI policy after the midterms: what changes in 2027 and what doesn't

For two years, US AI policy has followed one pattern: the White House sets direction by executive order, Congress debates without passing a broad AI law, and the states write the rules that actually bind. The Nov. 3, 2026 elections changed who controls Congress. They have not changed that pattern yet. Here is where things stand, and what leaders putting AI to work should do now.

What changed on Nov. 3

Democrats won majorities in both the House and Senate, and take control in January 2027. The IAPP notes that committee gavels bring subpoena power, hearing agendas and markup authority, with AI governance and accountability, federal privacy legislation, children's online safety and FTC oversight among the likely priorities.

AI was on the ballot in a way it had not been before. In an NBC News exit poll, 56% of voters said AI and data centers were a crucial or major factor in their vote for Congress, as TIME reported. Local fights over data center siting and power bills are tracked in our AI Power and the Grid hub.

Incoming majorities have already sketched an agenda. Reuters reports that Democratic lawmakers plan bills to protect consumers from electricity costs tied to data centers, require independent testing of advanced AI systems, and clarify when AI companies are liable for harm. Any bill still needs the president's signature or a two-thirds vote in both chambers to override a veto, and some analysts, including Russell Investments, expect divided government to limit major federal changes through 2028. Expect hearings and oversight to move faster than new statutes.

Washington steers by executive action

Federal AI policy runs mainly through executive orders and voluntary commitments, not statute:

• A national framework push. The December 2025 order Ensuring a National Policy Framework for Artificial Intelligence set a goal of one national standard, created a Justice Department AI Litigation Task Force to challenge state AI laws, asked the FCC and FTC to consider federal standards that would preempt conflicting state rules, and tied some federal grant funding to state AI laws. It leaves room for state laws on child safety, data center infrastructure and state government use of AI.
• Voluntary pre-release testing. A June 2, 2026 order invites developers of designated frontier models to give the government up to 30 days of access before broader release, sets up an AI cybersecurity clearinghouse, and states that it creates no licensing or preclearance requirement.
• Super Intelligence. On Sept. 29, 2026, six companies (Anthropic, Google, Meta, Nvidia, OpenAI and xAI) signed the voluntary White House Accord on Super Intelligence, committing to internal controls, outside auditors and board-level review, with no penalties for noncompliance. A companion order directs agencies to say "Super Intelligence" instead of "artificial intelligence" in official communications, and a White House task force chaired by the Director of National Intelligence adopted its charter in October with 120 days to report.

None of these create binding duties for private companies. They shape federal buying, federal security work and the fight over who sets the rules.

States are writing the binding rules

States kept legislating through the preemption push. The Transparency Coalition counted 115 AI-related laws enacted or passed in 30 states between Jan. 1 and Oct. 15, 2026, about 58% more than in 2025. Chatbot safety and child protections, limits on insurers' use of AI in coverage decisions, deepfakes and AI in education led the list.

A few laws matter most for organizations that operate in more than one state:

• California's frontier AI transparency law, SB 53, has been in effect since Jan. 1, 2026, alongside other state laws already on the books.
• Texas's Responsible Artificial Intelligence Governance Act took effect Jan. 1, 2026, covering state agencies' use of AI and certain prohibited uses.
• Colorado replaced its 2024 AI Act with SB 26-189, a narrower law for automated decision-making technology used in consequential decisions, effective Jan. 1, 2027. Deployers must give notice, explain adverse outcomes, and offer data correction and human review.
• New York's RAISE Act, as amended in 2026, sets safety and incident-reporting duties for the largest frontier model developers starting in 2027.

Preemption is still unresolved

No federal statute preempts state AI laws today. The Senate voted 99-1 in 2025 to strip a proposed 10-year moratorium on state AI laws from the budget bill, and a bipartisan House discussion draft released in June 2026 has not become a bill. In court, the Justice Department joined xAI's challenge to Colorado's original AI Act in April 2026; enforcement was paused while the state rewrote the law. State laws remain enforceable unless a court, Congress or the state itself says otherwise.

Whether the new Congress takes up preemption, and on what terms, is one of the biggest open questions for 2027.

Enforcement is moving faster than legislation

State attorneys general are using the consumer protection laws they already have. Eight states sued OpenAI on Oct. 14, 2026 over ChatGPT's safety claims and children's data, and a bipartisan group of 26 attorneys general asked Congress in September to set federal AI safety rules. Several incoming attorneys general have named AI harms and algorithmic pricing as 2027 priorities.

What leaders should do now

1. Comply with the state laws that apply to you. Executive orders do not displace state statutes. Plan for California, Colorado, New York and Texas obligations on their own timelines, and build to the strictest one you face.
2. Inventory AI used in consequential decisions. Hiring, lending, insurance, housing, health care and education draw the most rules. Know where AI touches those decisions, and be ready to give notice, explain outcomes and offer human review.
3. Check what you say about AI. Enforcement so far turns on claims: how safe, accurate or appropriate an AI product is said to be. Make sure marketing and disclosures match reality.
4. Mark January 2027. Colorado's new law, the new Congress and incoming attorneys general all arrive within weeks of each other.
5. Take part. Rulemakings like Colorado's invite public comment, and 2027 hearings will shape federal bills. For public sector leaders, procurement terms are policy too.

Follow it with ETA

ETA tracks AI policy news every day on the AI and Technology Policy hub, with a short take on what each story means for leaders. Policy, government and industry leaders meet at the US AI Congress, March 24 to 25, 2027 at the National Press Club in Washington, D.C.

This article is a nonpartisan overview for leaders adopting AI. It is not legal advice.