You don't get to wait for the courts. Your Q3 filings are due either way.
That is the operating reality for every U.S. carrier in 2026 after Executive Order 14365, signed December 11, 2025, opened a federal-state preemption fight over AI regulation that the NAIC publicly opposed within days. The EO does not specifically exempt insurance, and the NAIC is asking the administration to "reconsider this Executive Order and, at a minimum, affirm state regulation of AI in the business of insurance." Until that ask is answered, or until a court resolves the preemption question, carriers operate under two regimes at the same time. State insurance departments are continuing to issue AI bulletins, demand AI inventories, and run examinations. The federal government is asserting an authority that, on its face, threatens those activities. Both sides are right that the other side is doing something. Neither side has yielded.
Careful reading of statutory language will not resolve this problem. The resolution comes through litigation, agency rulemaking, or the next administration's reversal of position. Carriers do not get to pause their compliance calendars while that resolution unfolds. The Q3 filings, the pilot tool responses, and the state-level governance attestations are due on the schedule the states set, regardless of what is happening in federal court.
Here is what is actually happening, and what to do about it on a calendar that does not move.
What EO 14365 Preempts and What It Does Not Carve Out
The Executive Order, by the NAIC's own characterization, asserts a broad federal authority over state legislative and regulatory activity related to artificial intelligence. The NAIC's December 16, 2025 statement is unusually direct: the order "could disrupt well-established processes that ensure fairness and transparency in insurance markets and safeguard consumers from unfair or discriminatory practices," and "could implicate routine analytical tools insurers use every day."
The phrase "routine analytical tools" deserves attention. The NAIC is signaling that the EO's preemption claim reaches well beyond the recent generation of AI model bulletins, into traditional insurance analytics, scoring models, underwriting algorithms, and rate-setting tools that have been the subject of state regulation for decades. If that reading holds in court, the scope of preemption is much wider than a debate about generative AI policy. It reaches the tools that carriers and regulators have been negotiating over since the early 2000s.
The order does not contain a stated insurance carve-out. There is no language affirming the McCarran-Ferguson Act's reverse-preemption framework, which has historically protected state insurance regulation from federal statutes that do not specifically address insurance. Carriers and their counsel will argue, correctly, that McCarran-Ferguson should apply. The NAIC will argue the same thing. Whether the courts agree is the open question.
The honest reading is that the legal posture is unsettled and will stay unsettled for at least the next twelve to eighteen months. A carrier that builds its 2026 compliance plan on the assumption that EO 14365 will preempt state AI authority is making a bet against the historical baseline of insurance federalism. A carrier that builds its plan on the assumption that EO 14365 is irrelevant is making a bet against the explicit text of an executive order signed by the President. Neither bet is sound. The defensible posture is to comply with both regimes and let the lawyers argue about which one binds.
Why McCarran-Ferguson Is the Carriers' Ambiguity to Manage
McCarran-Ferguson is the doctrinal anchor most carrier counsel will reach for first. The 1945 Act provides that no federal statute shall be construed to "invalidate, impair, or supersede" state insurance regulation unless it specifically relates to the business of insurance. Federal courts have applied that framework for eighty years to a wide range of preemption questions, generally with deference to state insurance regulators on matters that touch underwriting, claims handling, and market conduct.
The application of McCarran-Ferguson to an executive order, rather than to a federal statute, is the first complication. McCarran-Ferguson speaks of "Acts of Congress." Executive orders are not Acts of Congress. The argument that an executive order issued without a specific congressional grant of authority over insurance cannot displace state regulation is a strong one, but it has not been litigated under EO 14365 yet. Until it has, the doctrine is a defense, not a settlement.
The second complication is that EO 14365's reach is presumably grounded in a broader federal authority claim, possibly under the Commerce Clause or under a national-security theory. Each of those theories has its own preemption analysis, and each one interacts with McCarran-Ferguson differently. Carrier counsel should expect the federal government, if challenged, to argue that the AI tools the EO targets are sufficiently national in their effects that traditional state-by-state regulation is inadequate. The counter-argument, which the NAIC has begun assembling, is that state regulators have demonstrated effective AI oversight through model bulletins adopted in over half the states.
The carriers most exposed to this ambiguity are not the largest national writers, who can absorb dual-track compliance costs. They are the regional and mid-sized carriers writing in states that have been most aggressive on AI bulletins, where the cost of state-level compliance is meaningful and the benefit of federal preemption, if it materializes, would be material. Those carriers face the worst version of the dual-regime problem: real cost to comply with state requirements, real uncertainty about whether the federal preemption claim will hold, and limited resources to hedge both.
The pragmatic approach for carrier counsel is to document, contemporaneously, the costs of dual compliance and the points at which state and federal requirements actually conflict. That documentation will matter if the preemption issue is litigated, and it will matter more if Congress or the courts eventually have to draw lines. Carriers without that documentation will be arguing from memory in two years. Carriers with it will be arguing from a record.
The 2026 Compliance Calendar That Does Not Move
Set the federal-state fight aside for a moment. The deadlines on the state side are not waiting for it.
The NAIC AI Systems Evaluation Tool pilot is running through September 2026 in twelve states. Carriers in California, Colorado, Connecticut, Florida, Iowa, Louisiana, Maryland, Pennsylvania, Rhode Island, Vermont, Virginia, and Wisconsin should expect inquiries and should treat the EO 14365 question as orthogonal to their pilot response. A carrier that refuses to respond, citing federal preemption, will face an examination posture problem from the state DOI well before any preemption litigation reaches a ruling. The defensible response is to cooperate with the pilot while preserving any preemption defenses through a clearly noted reservation of rights.
State AI bulletin compliance attestations remain on their existing schedules. The roughly two dozen states that adopted AI model bulletins before the EO have not withdrawn them. State market conduct examinations are continuing to incorporate AI inquiries into existing exam cycles. Q2 and Q3 rate filings that include algorithmic models continue to require the disclosures the state actuarial offices have asked for. None of those calendar items pause for the federal litigation.
The NAIC Fall National Meeting is in November 2026. Adoption of the AI Systems Evaluation Tool for nationwide use is on the agenda. Trade group attempts to delay or modify that adoption will continue, and EO 14365 will be cited in those attempts, but the adoption decision is a state regulator decision, not a federal one. The NAIC's December 16, 2025 statement signals where the state regulators are: defending their authority and accelerating, not pausing, their AI oversight work.
For carrier compliance leadership, the Q3 2026 governance playbook needs to assume the state regulatory baseline holds and add a federal-track work stream for the EO question. The work stream is small in volume, large in stakes: monitor the litigation, document dual-compliance costs, preserve preemption defenses through reservations of rights in regulatory responses, and coordinate with industry trade groups on amicus filings as preemption challenges reach the appellate courts.
There is a third dimension worth flagging, which is procurement. Carriers reviewing vendor AI contracts in 2026 should add an EO 14365 contingency clause to renewals and new agreements: language that allocates compliance cost and disclosure responsibility between carrier and vendor in the event that federal preemption either materializes or fails. Vendors are quietly shifting their own positions on which regime they believe will hold, and contract language signed today without that allocation will become a renegotiation problem in eighteen months. Carrier procurement teams should not assume their vendors are tracking the litigation closely. Most are not.
The carriers who set this up cleanly in Q2 will spend Q3 executing on a plan. The carriers who wait for the federal-state fight to resolve will spend Q3 trying to read tea leaves while their state filings come due. The first group will be visibly better prepared at year-end. That visibility, to regulators, to rating agencies, and to boards, is the operational dividend of treating EO 14365 as a planning input rather than a planning obstacle.
The fight over AI governance authority is real, and it is not going to be settled in 2026. The compliance calendar that runs alongside it is also real, and it is going to be settled every quarter, on the schedule the states already published. The carriers who recognize that the second calendar binds, regardless of what happens to the first, are the ones who will be compliant when the year ends.
