State AGs Push Back on FCC AI Preemption Effort
California Attorney General Rob Bonta and Colorado Attorney General Philip Weiser, joined by 23 bipartisan attorneys general, have filed a forceful reply comment with the Federal Communications Commission (FCC) opposing any attempt to preempt state artificial intelligence (AI) laws under federal telecommunications authority.
The letter responds to the FCC’s September Notice of Inquiry (NOI), which floated the idea that state and local AI laws could be preempted under Section 253(a) of the Telecommunications Act of 1996 if they are deemed to “prohibit or effectively prohibit” wireline telecommunications services. The attorneys general reject this premise outright.
The FCC’s notice of inquiry lays the groundwork for the agency to attempt to preempt state AI laws in future proceedings. The attorneys general argue this premise is legally flawed, procedurally defective, and constitutionally suspect.
AI as a Barrier to Telecom Deployment
At the center of the dispute is Paragraph 53 of the FCC’s NOI, which asks whether state laws “seeking to govern or limit uses of AI” could unlawfully interfere with telecommunications services as AI becomes more embedded in network operations, customer service, fraud prevention, routing, and optimization.
Industry commenters supporting preemption have urged the FCC to treat a wide range of AI-enabled software tools as functionally inseparable from telecom service provision, opening the door to broad displacement of state oversight.
The states sharply reject that framing.
States’ Core Legal Objections
The reply comments advance several interlocking arguments:
1. AI Is an Information Service — Not Telecom
The states emphasize that AI is best understood as software and information processing, squarely outside the FCC’s Title II jurisdiction. AI systems generate, analyze, rank, predict, and transform information; they do not transmit communications “without change in form or content,” as required to qualify as telecommunications under federal law.
Because the FCC lacks authority to regulate information services, it likewise lacks authority to preempt state laws governing AI. Courts have repeatedly held that an agency cannot preempt where it has no underlying regulatory power.
2. Section 253 Cannot Bear This Weight
Section 253(a) has historically been applied narrowly, only where state or local laws materially inhibit market entry or competition by telecommunications providers. The states argue that consumer protection, civil rights, housing, healthcare, election integrity, and privacy laws addressing AI do not remotely resemble barriers to telecom deployment.
Treating any AI-related state law as a potential Section 253 violation, they warn, would effectively grant the FCC roving authority over the entire economy.
3. Major Questions & Tenth Amendment Problems
The attorneys general invoke the major questions doctrine, noting that Congress has never clearly authorized the FCC to preempt state regulation of AI. Any attempt to do so through an NOI, without explicit statutory authorization, would almost certainly fail judicial review.
They further argue that sweeping preemption would intrude on core state police powers protected by the Tenth Amendment, including regulation of professional conduct, housing markets, elections, consumer fraud, and privacy.
4. APA Failures: Too Vague to Act
Procedurally, the states contend the NOI is fatally vague under the Administrative Procedure Act. It fails to define “artificial intelligence,” identify specific state laws at issue, or demonstrate any real-world impact on telecom deployment. Without meaningful notice, any resulting FCC action would be arbitrary and capricious.
Why States Say This Matters
The filing underscores that AI is already embedded across daily life from rent-setting algorithms and healthcare diagnostics to deepfakes, automated hiring, pricing tools, and content moderation. States are often the first to receive consumer complaints, identify harms, and develop targeted safeguards.
The attorneys general stress that states have long functioned as regulatory laboratories, experimenting with disclosures, opt-outs, identity protections, and sector-specific AI rules that respond to real-world harms faster than federal agencies typically can.
As AG Bonta notes, states have proven they can “both nurture evolving innovation and protect our residents,” rejecting the notion that AI governance and economic growth are mutually exclusive.
LegisLogiq Outlook
This filing crystallizes a growing national fault line in AI governance. As federal agencies explore centralized AI control, often by stretching legacy statutes, states are asserting their role as regulators of first resort.
“I’ve said it many times before and I’ll say it again: I strongly oppose any effort to block states from developing and enforcing common-sense AI regulation.”
Earlier this month, the Trump Administration released an executive order placing independent regulatory agencies at the center of the preemption effort. The FCC is directed to initiate a proceeding to consider a federal AI reporting and disclosure standard that would override state requirements. If adopted, this could become the first explicit federal AI transparency framework with nationwide preemptive effect.
Last month, Attorney General Bonta joined a bipartisan coalition of 36 state attorneys general in sending a letter to Congress opposing a proposed provision in the National Defense Authorization Act that would preempt state laws addressing the risks of AI. In the letter, the attorneys general argue that, while artificial intelligence promises to be a transformative technology in multiple fields, it also poses significant risks, especially to children, and that states must be empowered to utilize existing laws and formulate new approaches to mitigate potential harms associated with artificial intelligence
The FCC’s NOI reflects a broader Trump Administration posture favoring aggressive federal preemption of state technology laws. But without clear congressional action establishing a federal AI framework with explicit preemptive scope, agency-driven efforts to displace state AI laws are likely to face intense legal resistance and skeptical courts.
AI preemption is rapidly becoming the next major battleground in U.S. federalism and the states are not backing down.
