If you file a brief in Florida, you have to certify that any AI-generated content has been independently verified by a human. Texas has the same requirement. So do Minnesota, Massachusetts, and New York.

Iowa does not. Not yet.

But five states have moved from guidance to formal rules in under 18 months, and California is actively drafting its own version. “Iowa hasn’t required it” is a temporary statement. The question worth asking is whether your firm will be ready when it does.

What these rules actually require

The specifics vary by state, but the core obligation is the same: if AI tools were used in preparing a court filing, the attorney must certify that the output was independently verified against primary sources.

Florida was first, effective June 15, 2025. The state’s standing order requires attorneys to disclose AI use and certify that all citations and legal assertions generated by AI have been verified by a licensed attorney. It was a direct response to a string of sanctions in federal courts over fabricated citations.

Texas followed with a similar order covering state courts. The certification language specifically targets “generative artificial intelligence” and requires that attorneys take personal responsibility for every factual and legal assertion, regardless of whether a human or a machine produced the first draft.

Minnesota and Massachusetts adopted rules through their respective supreme courts. Both focus on citation verification and require disclosure when AI tools substantially contributed to brief preparation.

New York went live June 1, 2026. The rule applies across state courts and requires attorneys to certify that they have reviewed and verified any AI-assisted content in their filings. New York’s rule is notable because it comes from the largest state court system in the country, and because multiple New York federal judges had already been imposing their own AI disclosure requirements for over a year before the state acted.

The sanctions are real

Courts are enforcing these requirements, and the penalties have teeth.

A New York attorney was fined $2,500 in an SDNY case for filing AI-generated content without verification. The court described the filings as “outrageously unprofessional.”

The Sixth Circuit imposed $30,000 in sanctions against attorneys who submitted briefs with fabricated citations generated by AI. The court was direct: the tools are not the problem. The failure to verify is the problem.

In Mississippi, an attorney’s case was removed from the court’s docket entirely after AI-generated fabrications were discovered in filings. Not a fine or a warning. The case was gone.

And in one of the highest-profile incidents of 2026, Sullivan and Cromwell, an Am Law 10 firm with mandatory AI training, admitted to filing roughly 40 AI-hallucinated citations in federal bankruptcy court. They had policies. They had training modules with completion tracking. None of it prevented the error because nobody actually ran the verification step on the work that went out the door.

Why this matters for Iowa firms

Iowa’s bar hasn’t adopted a filing certification rule. The Iowa State Bar Association established an AI Committee in December 2025 and ran a six-session AI Training Series through spring 2026, but formal rulemaking hasn’t followed.

That doesn’t mean Iowa firms are off the hook.

First, any Iowa attorney who files in federal court may already be subject to individual judges’ AI standing orders. There’s no uniform federal rule requiring AI disclosure. The Judicial Conference hasn’t amended the Federal Rules of Civil Procedure to address it. Instead, dozens of individual federal judges have issued their own standing orders, and some districts have adopted local rules. The requirements depend on which judge you draw, which is arguably worse than a clear rule because it means checking standing orders case by case. If your firm handles any federal litigation, look up the standing orders for your assigned judges before your next filing.

Second, if your firm does business across state lines (and many Iowa firms do, particularly in the Omaha-Council Bluffs metro or with clients in Illinois, Minnesota, or Missouri), you may already be filing in jurisdictions with certification requirements. Minnesota’s rule is active right now.

Third, judges have tools to catch this. Thomson Reuters provides CoCounsel Judicial Edition to courts, giving judges AI-powered brief-checking capability. Iowa judges have access to tools that can flag AI-generated errors in your filings. The asymmetry matters: you might not know whether AI hallucinated a citation, but the judge’s software might.

What to document now

You don’t need to wait for Iowa to adopt a rule to get your house in order. The firms that come through this cleanly will be the ones with processes already running when the requirement arrives.

A written AI use policy. ABA Formal Opinion 512 (July 2024) already says the duty of competence extends to AI tools and the duty of supervision covers staff AI use. If your firm uses AI for anything touching client matters, put it in writing: which tools are approved, how output gets verified, and who signs off. Our post on AI acceptable use policies covers what belongs in one.

A verification workflow. The policy is only useful if there’s a concrete process behind it. Who checks AI-generated citations? How? Is there a sign-off step before filing? Sullivan and Cromwell had the policy but not the workflow. The verification step has to be built into production, not bolted on after the fact.

Client disclosure language. Your engagement letter should address AI. Both your firm’s use of it and your clients’ use of it. The Heppner ruling (S.D.N.Y., February 2026) stripped privilege from consumer AI conversations. If your clients are putting your legal advice into ChatGPT, that’s discoverable. Your engagement letter should say so.

An AI tools inventory. Know what’s actually in your environment. Which attorneys are using which tools? Is anyone on staff running client work through consumer AI? Shadow AI is a real problem in law firms, and you can’t govern what you don’t know about.

Where this is headed

Five states have formal certification requirements. California is drafting its version. Federal judges across the country have been writing their own standing orders for over a year. ABA Opinion 512 set the ethical baseline back in 2024. And the sanctions keep getting bigger: from $2,500 fines to $30,000 penalties to outright case removal.

Iowa firms that build their verification processes now won’t have to scramble when the rule arrives. They’ll already be filing clean.

If your firm needs help creating an AI use policy or building a verification workflow that actually gets followed, reach out.