0

Attempts to halt or preempt state AI rules with will be determined by today’s deliverables from Commerce Dept. and FTC

The March 11 deadline has arrived – the critical “action date” for federal agencies like the Department of Commerce (DOC) and Federal Trade Commission (FTC) to implement core directives under Executive Order 14365, “Ensuring a National Policy Framework for Artificial Intelligence.” The EO came after a wave of state-level AI regulations were issued to address a regulatory vacuum about AI-related harms like bias, deepfakes, chatbot disclosures, automated decision-making, and dubious bot-related hiring practices.

While the EO mentions only Colorado’s AI Act, the Trump administration has since mentioned California (SB 53 and AB 2013) and New York’s (RAISE Act). Since, other states have enacted similar laws, including Texas (TRAIGA Act), Utah (AI Policy Act), Illinois (AIVIA), and Maryland (H.B. 1202).

The EO aims to halt or preempt state regulations deemed “onerous” by federal agencies, which means those that may impede the Trump administration’s policy to achieve “global AI dominance through a minimally burdensome national policy framework for AI.” As put forth by the White House, the order aims to reduce the “patchwork” of 50 different state regulatory regimes into a single, minimally burdensome federal standard.      

In yesterday’s Mondaq blog by Joseph Cahill of Baker Botts law firm, “this evaluation will identify specific state laws that the administration considers problematic and may refer to the DOJ Task Force.”

As such, this is the day – 90 days after the EO was signed – by which the DOC must issue a report on state AI laws for which it feels a federal preemption of state regulation is necessary.

As explained by Cahill in his blog, “The Executive Order does not itself preempt, repeal, or invalidate any state AI law. Instead, it directs a sequenced series of federal agency actions that could result in litigation challenging state AI laws, new agency guidance on the application of existing federal statutes, and conditions on federal funding.”

To that end, new levers have been created for the DOJ by Attorney General Pam Bondi, who has created the AI Litigation Task Force, a unit that in the EO will consult with David Sacks, AI and Crypto Czar and Howard Lutnick, the Secretary of Commerce, on state laws it feels “unconstitutionally burden interstate commerce or conflicts with federal regulations.”

By asserting federal control over the “patchwork” of state AI laws, this action attempts to shift the balance of federal and state lawmaking, triggering a sequence of legal and financial challenges in an attempt to force national uniformity. The first step is identifying state that are “onerous,” and then by providing a framework of agency guidance and funding conditions meant to make federal monies, such as that for the $21 billion Broadband Equity Access and Deployment Program, conditional on alignment with federal AI policies.

In today’s report, the Department of Commerce will identify states whose laws

• Force AI models to “alter their truthful outputs.”
• Compel disclosures that might violate the First Amendment.

The second deliverable due today is a policy statement from the Federal Trade Commission on the application of Section 5 of the Federal Trade Commission Act (15 U.S.C. § 45(a), which declares “unfair or deceptive acts or practices” (consumer protection) and “unfair methods of competition” (antitrust) in or affecting commerce to be unlawful. It empowers the FTC to act against practices beyond the scope of traditional antitrust laws, such as non-compete agreements or deceptive marketing.

The Trump administration is seeking from the FTC Chairman an explanation of the circumstances under which state laws “require alterations to the truthful outputs of AI models,” as preempted by Section 5 of the FTC Act.

When it comes to AI models, Section 5 pertains to AI developers and how they may adjust model outputs to mitigate bias, which in the Trump administration’s view, might compel the production of outputs that are “deceptive” under federal law.

What the impact of today’s reports by the Dept. of Commerce and FTC remains to be seen, but at a minimum, all AI-related entities should become familiar with their respective state laws – those that have been passed and those that might be.

For now, state laws that have taken effect “are enforceable absent court action,” said Cahill, who urges companies to “map which of their AI-related obligations could be affected by a federal challenge, prepare to adapt compliance programs if preliminary relief is granted in specific jurisdictions, and monitor three developments closely: (1) the breadth of the Commerce Department’s evaluation when it is published; (2) the legal theory the FTC articulates in its policy statement; and (3) whether the DOJ Task Force moves quickly from the Commerce referrals to actual litigation.”

In short, the DOC and FTC cannot preempt state law, but they can get the DOJ to challenge state laws  in court, or through separate actions by the FTC or DOC.

Without the backing of Congress, the EO’s attempt at “preemption by executive action” has come under legal scrutiny, but it does manage to create uncertainty for states and AI-related entities in those states when it comes to AI-related harms and the uncertainties of the existing regulatory environment at a state and federal level.

RCRTech will continue to report as the situation evolves.