AI rules, Arizona‐style

Not bad, Arizona Legislature. Not bad at all.

We were encouraged when we saw the Arizona House created an AI-focused committee earlier this month. Now, we’re seeing bills coming from lawmakers that actually make a lot of sense, in our humble opinion.

None of the bills is perfect, of course, but each one is geared toward addressing real-world problems, from kids using chatbots to the complexity of divorce proceedings.

Without further ado, let’s start with the type of bill that lawmakers love to pass: the ones that give them all the power!

This bill comes from the Legislature’s newly crowned AI chief, Republican Rep. Justin Wilmeth, the chair of the AI and Innovation Committee in the House.

Under the bill, the state Department of Administration would direct agencies to find new ways to use AI, such as using AI to cut red tape or streamline procurement.

Agencies also could look for regulations that “unnecessarily restrict” AI innovation and propose repealing them.

But that’s as far as agencies would be able to take it. The regulatory steering wheel would stay firmly in the grip of the Legislature.

Any new AI-related rules that an agency comes up with wouldn’t take effect unless lawmakers sign off, even in the case of emergency or temporary measures.

Republican Rep. Alexander Kolodin is trying to turn AI literacy into a statewide summer program, rather than a STEM-kid niche.

His bill would create the Arizona Artificial Intelligence Education Program within the Department of Education.

It would be a voluntary course open to all Arizonans, not just K-12 students. Participants would learn about digital hygiene and civic integrity.

The program also would offer some help for entrepreneurs. The Office of Economic Opportunity would draft a small-business curriculum, recruit volunteer instructors and source locations.

The bill is designed to be integrated into public educational institutions, from K-12 schools to community colleges to state universities. Officials at those institutions could award credit or lend space to the program for free, while plugging the AI education program into their pre-existing programs.

If it’s well‑run, this could double as both workforce development and a civic‑resilience play against AI‑driven misinformation.

Republican Rep. Teresa Martinez is testing AI in one of the most sensitive arenas: family law.

If both parties in a divorce agree to use “artificial intelligence‑assisted arbitration,” then they could let ChatGPT help hammer out the details.

The catch is they wouldn’t be able to use AI arbitration if they share minor children. And either party could withdraw consent at any time before a binding determination comes out.

Martinez’s bill also leaves the door open for a superior court judge to re-decide any award that came from the AI-arbitration.

In practice, this could treat AI as a turbocharged calculator for settlements, rather than an autonomous judge, and keep the human courts fully in charge.

This bill from Republican Rep. Tony Rivero zeroes in on conversational AI used by minors and reads like a direct response to scandals around unsafe AI companions.

Any operator of publicly accessible conversational AI tools would have to clearly tell minors that they’re talking to AI, not a human.

If the user is a minor, the operator also would have to build safeguards against sexual content, deceptive “I’m human” simulation, emotional manipulation, and addictive engagement features.

The bill would bar operators from marketing these services as professional mental or behavioral health care. The bill would also require the AI tools to route suicidal or self‑harm prompts to appropriate services.

The bill would mandate privacy tools for minors and make account-management tools available to parents or guardians.

The attorney general would be in charge of enforcing the new rules, which could lead to civil penalties up to $1,000 per violation, with a cap of $500,000 per operator.

The bill would take effect in October 2027, which would give the AI industry time to re-architect safeguards for kids.

Kolodin’s second AI-related bill would basically grant the same legal privilege you expect when you talk to your doctor or lawyer to your conversations with AI chatbots.

That could come in handy for people who use AI to draft or refine questions for a lawyer, therapist or financial advisor.

The bill itself is short and sweet: “A person’s communication with an artificial intelligence is privileged if the person would have been entitled to privileged communication had the person sought the advice from a human professional.”

At the same time, the bill would nudge courts to grapple with how far privilege can really stretch when third‑party tech infrastructure is in the middle.

These bills are just the opening salvo in what likely will be months of debate at the Legislature. They’ll have to be heard in the AI and Innovation Committee, (which meets for the first time today!) and any other committee they get assigned to in the House.

Then they’ll have to go through the same process in the Senate, before they could be sent to Gov. Katie Hobbs’ desk to be signed into law or vetoed.

We’ll keep track of it all for you, of course. We’re just as intrigued as you are!

But if you want to follow along on your own, feel free to use our tracking list.

And if you’re really into tracking bills, check out Skywolf, our legislative intelligence service. It’ll blow you away.