AI rules by state · National baseline
AI rules for lawyers in states without specific guidance
Quick answer
Many states have not issued their own AI ethics opinion. In those states, the same core duties still apply: competence, candor to the tribunal, confidentiality, and supervision. No state permits a lawyer to file unverified AI output. The ABA's Formal Opinion 512 supplies the national framework that these states follow, and the duty that matters most everywhere is to verify what an AI tool produces.
The national framework
ABA Formal Opinion 512
Where a state has not issued its own AI guidance, the American Bar Association's Formal Opinion 512 is the common reference. Issued in July 2024, it was the ABA's first formal ethics opinion on generative AI.
The opinion applies the existing Model Rules to AI use. It addresses competence, the protection of client confidentiality, communication with clients, candor toward the tribunal, the supervision of work, and reasonable fees. The throughline matches the states that have issued their own opinions: the lawyer remains responsible for the work, and the duty to verify AI output cannot be delegated to the tool.
Source: ABA Formal Opinion 512, Generative Artificial Intelligence Tools (PDF)
Three approaches to AI disclosure
Why the answer changes by jurisdiction
Courts and bars fall into three broad camps on whether a lawyer must disclose AI use. Knowing which camp a jurisdiction is in tells you what to expect, even where a state has issued little guidance of its own.
Require disclosure or certification
Some courts require a signed statement about AI use in every filing. The first federal example was Judge Brantley Starr in the Northern District of Texas, whose order requires each filer to certify either that no generative AI was used, or that any AI-drafted text was checked for accuracy by a human.
Rely on existing duties
Most states issue no special disclosure rule. The existing duties of competence and candor to the tribunal apply, and the lawyer must independently verify AI output before relying on it.
Advisory guidance, or nothing specific
Many states have only general guidance, or none specific to AI. Lawyers there look to the national ABA framework, and the same duties still apply.
Sources: ABA Formal Opinion 512, Judge Starr standing order (N.D. Tex.)
The pattern across states
No state lets you skip verification
Two patterns hold across the states that have addressed AI, and they are the safest assumptions in a state that has not. First, no state mandates a blanket disclosure of AI use to a court or a client. Second, every state requires the lawyer to verify AI output before relying on it.
The consequences of skipping verification are national, not local. Courts across the country have sanctioned lawyers for filing AI-fabricated citations, starting with Mata v. Avianca in 2023. A state without its own opinion is not a safe harbor. The same duties, and the same risks, apply.
Source: Mata v. Avianca, Inc., S.D.N.Y. sanctions order docket (CourtListener)
The competence duty everywhere
Most states follow ABA Model Rule 1.1
Most states have adopted a version of ABA Model Rule 1.1, the duty of competence, including Comment 8, which directs lawyers to stay abreast of the benefits and risks associated with relevant technology. That duty covers AI even in states that have issued no AI-specific opinion.
In practice this means a lawyer must understand a tool's limits, including its tendency to fabricate, and must review and correct its output through independent judgment.
Source: ABA Model Rule 1.1, Comment 8 (Maintaining Competence)
How to stay compliant
A practical checklist for any state
Review the assigned judge's standing order and the court's local rules at the start of every matter. AI disclosure obligations generally sit at the court and judge level rather than in a single statewide rule.
Verify every citation and quotation before filing, and read the underlying authority. Most AI sanctions have resulted from citations the lawyer never personally verified.
Do not enter confidential client information into public AI tools without confirming how the tool handles data and, where there is material confidentiality risk, obtaining the client's informed consent.
Inform the client when AI use materially affects the matter, and honor any client instruction that limits AI use.
Treat AI competence as part of your duty under your state's version of the competence rule, which in most states follows ABA Model Rule 1.1.
When evaluating tools, look for citation grounding that ties output to verifiable authority, a vendor commitment not to train on client data, and audit trails. These reduce risk, but they do not replace the lawyer's own review.
States this page covers
States without their own AI opinion
As of June 2026, the following states had not published their own AI ethics opinion, AI court rule, or in-state AI-hallucination sanction. Lawyers in these states rely on the national framework above.
A few states have a bar task force or early activity underway, but no binding AI opinion as of June 2026. The same national duties apply while their guidance develops.
Frequently asked
AI rules where a state has no opinion: common questions
This page is general information, not legal advice. State guidance changes quickly, and the obligations that apply to your matter depend on your court, your judge, and your facts. Verify the current rules with your state bar and the assigned court before relying on anything here. Last verified June 2026.
Built for lawyers who have to verify everything
Irys grounds its output in verifiable authority, keeps client data out of training, and maintains a record of the work. See how it fits your practice.
Book a Demo