Coalition of State AGs Push Back Against FCC Proposal Seeking to Preempt State AI Laws

In reply comments filed in the FCC’s “Build America: Eliminating Barriers to Wireline Deployments” proceeding, the states contend that the Commission should abandon any declaratory ruling that purports to preempt state and local laws governing or limiting uses of AI. The comments respond directly to industry filings urging the FCC to interpret Section 253 of the Communications Act as authorizing broad federal preemption of state AI requirements that could allegedly interfere with telecommunications services.

In a Notice of Inquiry (NOI) issued in September, the FCC asked whether state or local laws “seeking to govern or limit uses of AI” might effectively prohibit the provision of telecommunications services and therefore be subject to preemption under Section 253 of the Communications Act. Impetus for the NOI came from the Trump administration’s AI Action Plan issued in July,

Supporters of the FCC’s proposed approach, including telecommunications providers and trade associations, argue that AI is increasingly integrated into network management, customer service, fraud detection and other operational functions, making state AI regulation a potential barrier to communications infrastructure deployment.

The state attorneys general reject that premise outright. Their central objection is that “artificial intelligence” is an extraordinarily broad and undefined concept that encompasses a wide range of software-based information services far beyond the FCC’s jurisdiction. The filing emphasizes that AI is not limited to generative models or chatbots, but includes longstanding forms of machine learning used in healthcare, e-commerce, transportation, finance and even routine legal filings. Treating AI as a category subject to FCC preemption, the states argue, would amount to federal oversight of software across the entire economy, a result Congress has never authorized

Legally, the states maintain that Section 253 applies only to state or local measures that materially inhibit the provision of telecommunications services, defined as the transmission of information without change in form or content. AI-enabled tools, by contrast, transform, process and generate information and therefore fall squarely within the statutory definition of “information services,” an area the FCC lacks authority to regulate or preempt.

Related: Federal AI Strategy Raises Compliance Stakes for Banks and Big Tech 

Citing recent appellate court decisions, the attorneys general argue that where the Commission lacks regulatory jurisdiction, it also lacks preemptive power, making any AI-focused declaratory ruling unlawful from the outset.

The filing also invokes the Supreme Court’s major questions doctrine and federalism principles. Preempting broad categories of state AI law, the states argue, would represent a basic and fundamental shift in the statutory scheme governing communications, one that would require unmistakably clear congressional authorization. No such authorization exists, and the FCC cannot supply it through interpretation alone.

The comments further warn that sweeping preemption would raise serious Tenth Amendment concerns by intruding on traditional state responsibilities such as consumer protection, professional discipline, election integrity and housing regulation.

The filing highlights a range of legitimate state interests that could be impaired by FCC action. They include laws addressing AI-generated deepfakes, disclosures when consumers interact with automated systems, algorithmic rent-setting, identity protection in synthetic media, and consumer rights to opt out of consequential automated decisions.

None of these measures, the states argue, is plausibly related to limiting telecommunications market entry or infrastructure deployment. At least two dozen states have passed or enacted laws addressing one or more of highlighted issues.

Finally, the filing claims the FCC’s process is deficient under the Administrative Procedure Act. The NOI contains no definition of AI, identifies no specific state laws subject to preemption, and points to no factual record showing actual barriers to telecommunications services. That lack of clarity, the states contend, deprives stakeholders of meaningful notice and would render any resulting declaratory ruling arbitrary and capricious if adopted.