AI rules in the US in 2026: Federal policy, state laws and what changes for you

The United States still has no single, comprehensive federal law governing artificial intelligence in 2026. Instead, the White House is pushing for one national standard that would override many state rules, while states such as California and Colorado keep passing their own laws on chatbots, hiring algorithms and the most powerful AI models.

For consumers, protections depend heavily on where you live, and they could shift again if Congress acts or federal challenges to state laws succeed.

Is there a federal AI law in the US?

Not yet. Federal policy runs mainly through executive orders and voluntary standards. The central document is Executive Order 14365, "Ensuring a National Policy Framework for Artificial Intelligence," which President Donald Trump signed on Dec. 11, 2025, according to the White House. It directs agencies to push back against state AI laws the administration considers burdensome, including through:

• A Justice Department task force dedicated to challenging state AI laws in court, on grounds such as interference with interstate commerce or federal preemption.
• A Commerce Department review to identify "onerous" state AI laws, including those the order says would force AI models to alter truthful outputs.
• Funding leverage, making states with such laws ineligible for certain federal broadband (BEAD) funds to the extent the law allows, and asking agencies to consider similar conditions on other grants.
• FCC and FTC actions on federal disclosure standards and deceptive AI practices.
• A legislative proposal for a federal framework preempting conflicting state laws.

The order also carves out areas it says should not be preempted, including state child safety protections, rules for data centers and AI computing infrastructure, and states' own government use of AI.

What is in the White House's national AI legislative framework?

On March 20, 2026, the White House released a set of legislative principles asking Congress to write national AI rules. According to the White House, the framework calls for:

• Uniform national rules that would replace a patchwork of state laws.
• Child safety measures, such as parental account controls and features to reduce exploitation and self-harm risks on platforms likely used by minors.
• Protections for creators, alongside a proposal to let AI systems make fair use of material they learn from.
• Free speech safeguards aimed at preventing AI from being used to censor lawful political expression.
• Energy protections, including keeping data center costs off household utility bills.

AP coverage in late September described California officials criticizing the absence of comprehensive federal AI regulation, a sign Congress had not yet acted on the framework.

What is the voluntary AI accord announced in September?

On Sept. 29, 2026, Trump said leaders of major technology companies had signed a voluntary agreement to police their own AI development, the Associated Press reported. The AP said signers included executives from Google, Meta, OpenAI and Nvidia, along with Elon Musk of xAI.

According to the AP, the accord centers on internal controls, independent external audits of those controls, board-level committees to review auditor findings and openness to future regulation. Trump described it as morally binding, but it is not a law.

Reaction was mixed: some academics quoted by the AP doubted self-policing would be enough, while another said it balances innovation and oversight.

What role does NIST play in AI rules?

The National Institute of Standards and Technology writes widely used technical guidance, but its main AI document is voluntary. NIST's AI Risk Management Framework, released in January 2023, is "intended for voluntary use" and helps organizations identify and manage AI risks.

NIST added a generative AI profile in July 2024 and in April 2026 published a concept note for a critical infrastructure profile. The agency says the framework is being revised under the White House AI Action Plan.

How many states have AI laws?

States enacted 109 AI laws in the first half of 2026 across 29 states, according to an analysis by NYU's Center on Technology Policy published in Tech Policy Press in July. The same analysis counted 14 new companion chatbot laws out of more than 100 bills introduced.

What changed in Colorado's AI law?

Colorado passed the first broad state law on algorithmic discrimination in 2024, but lawmakers substantially narrowed it in 2026. Gov. Jared Polis signed SB 26-189 on May 14, 2026, and the revised law takes effect Jan. 1, 2027, according to an analysis by the law firm Holland & Knight.

The new version focuses on "automated decision-making technology" used in consequential decisions about education, jobs, housing, lending, insurance, health care and government services, according to the state legislature's fiscal note. Holland & Knight reported that it drops the earlier duty of reasonable care against algorithmic discrimination, along with mandatory risk programs and annual impact assessments.

For Colorado residents, the law promises:

• Notice before such technology materially influences a major decision about you.
• An explanation after an adverse outcome.
• The right to correct inaccurate data and request human review, where commercially reasonable.

Only the attorney general can enforce it, and disclosure violations can carry civil penalties of up to $20,000 each, the fiscal note says. Separately, the fiscal note reports that a federal court in April 2026 barred enforcement of the original 2024 law pending rulemaking and a ruling in a lawsuit brought by xAI, and Holland & Knight reported that the Justice Department intervened in support of that challenge.

What AI laws has California passed?

California has been the most active state. Its 2025 Transparency in Frontier Artificial Intelligence Act, SB 53, requires developers of the most advanced models to publish safety frameworks, report certain critical safety incidents to the state and protect whistleblowers, according to Gov. Gavin Newsom's office.

In September 2026, Newsom signed another wave of measures, according to his office:

• Companion chatbots and children: SB 1119 requires crisis protocols for suicidal ideation, parental controls, alerts when a child disables safety settings, and independent child safety audits.
• Workplace AI: Bills including SB 947 on automated decision systems in employment and SB 951 on notice for technological displacement. The AP reported the package bars employers from relying on AI to decide to fire someone and from using AI to infer workers' emotions from biometric data.
• Transparency and provenance: Updates to the California AI Transparency Act (AB 2713 and SB 1000).

Newsom has cast the state's approach as filling a gap left by federal inaction, the AP reported.

What does this mean for you?

• Job applicants and workers in California and, from 2027, Colorado gain new notice and review rights when AI influences employment decisions.
• Parents in states with companion chatbot laws may see new safety settings and parental controls.
• Californians may notice more labels and provenance data on AI-generated content as the state's updated transparency laws take effect.
• Residents of other states depend largely on general consumer protection rules and on whatever Congress eventually passes.

The biggest open question is preemption. If Congress adopts a national standard or federal court challenges succeed, some state protections could be narrowed or blocked.

Frequently asked questions

Is AI regulated by the federal government in 2026?

There is no comprehensive federal AI statute. Federal policy relies on executive orders, agency actions and voluntary standards such as NIST's AI Risk Management Framework, while the White House has asked Congress to pass national rules.

Can the president stop states from passing AI laws?

Executive Order 14365 does not repeal state laws directly. It directs the Justice Department to challenge state laws in court, ties some federal funding to state AI policy and asks for a federal law that would preempt conflicting state rules, according to the White House.

When does Colorado's AI law take effect?

Colorado's rewritten law, SB 26-189, takes effect Jan. 1, 2027, according to Holland & Knight. The attorney general must adopt rules on disclosures and consumer rights by the same date.

What does California's SB 53 require?

SB 53 requires developers of frontier AI models to publicly disclose safety frameworks, report specified critical safety incidents to the state and protect whistleblowers. Newsom's office calls it the nation's first AI safety law.

Is the White House AI accord with tech companies legally binding?

No. The Associated Press reported that the September 2026 accord is voluntary, centered on internal controls and independent audits. It says turning those steps into laws or regulations may make sense over time.