Until now, most guidance on AI in legal practice has been advisory. The ABA published Formal Opinion 512 in 2024 saying the duty of competence extends to AI tools. State bars have issued ethics opinions. The message has been consistent: understand the tools you’re using, supervise staff, verify output.

But guidance and rules are different things. Guidance says you should. Rules say you must.

California is moving from “should” to “must.”

What California proposed

On May 5, the State Bar of California’s Standing Committee on Professional Responsibility and Conduct proposed a new comment to Rule 1.1 (competence). The key language:

When using any technology, including AI, a lawyer “must independently review, verify, and exercise professional judgment regarding any output generated by the technology that is used in connection with representing a client.”

That’s not a suggestion. It’s proposed rule language. If adopted, it becomes part of the competence standard that California attorneys are measured against in disciplinary proceedings.

The word “independently” is doing a lot of work in that sentence. It means you can’t just ask the AI if its own output is accurate. You have to check it against primary sources yourself. The Sullivan & Cromwell incident from April, where roughly 40 AI-hallucinated citations made it into federal court filings, is exactly the kind of failure this language is designed to prevent.

Why this matters outside California

California is the largest state bar in the country. When California adopts a rule change, other states pay attention. The pattern is well established: California or New York moves first, the ABA updates its model rules or commentary, and state bars across the country follow within a year or two.

Iowa hasn’t proposed similar language yet. But the Iowa Rules of Professional Conduct track the ABA Model Rules closely, and the ISBA has been running an AI training series throughout 2026 that covers exactly these issues. The May 2026 issue of the Iowa Lawyer Magazine includes an article on generative AI intellectual property risks. The ISBA session on May 15 is specifically about AI policy and vendor selection.

The infrastructure for an Iowa rule update is being built. The question is when, not if.

What this means for your practice

If you’re an Iowa attorney using AI tools today (and most are, whether officially or not), the practical takeaway is straightforward:

Verification needs to be a workflow step, not a reminder. “Trust but verify” doesn’t work if verification is optional. It has to be built into how work gets done. Someone other than the person who generated the AI output checks the citations, the case references, and the factual claims. Every time.

“I didn’t know it was wrong” isn’t going to work as a defense. The proposed California language requires lawyers to exercise professional judgment about AI output. That means you’re expected to recognize when something looks off, even if the AI presented it confidently. If you’re using AI for tasks you don’t understand well enough to evaluate the output, that’s a competence problem regardless of what state you practice in.

Your AI policy needs teeth. A policy that says “employees should verify AI output” is weaker than one that says “no AI-generated content is submitted to a court or sent to a client without independent verification by a licensed attorney, documented in [specific system].” If you need a starting point, we wrote about what a practical AI acceptable use policy looks like earlier this year. The firms that will be in the best position when their state bar adopts similar language are the ones that already operate this way.

The timeline

California’s proposal is in the comment period now. Even if it moves fast, it won’t change Iowa rules overnight. But the direction is clear. ABA Opinion 512 laid the groundwork. Sullivan & Cromwell showed what happens when policies alone aren’t enough. California is writing verification into the rules themselves.

Iowa attorneys who build verification into their workflows now won’t have to scramble when their own bar follows. And they won’t have to explain to a judge why they didn’t.

Artech Solutions helps Iowa law firms put AI governance policies in place, including verification workflows and acceptable use frameworks. If your firm is using AI tools and doesn’t have a written policy yet, let’s talk.