Arizona’s lawmakers are seeing AI for what it is: a tool to automate knowledge work.
Instead of casting AI as a world-ending technology, or hyping it up as the key to a leisure-filled utopia, lawmakers at last week’s hearing of the House AI committee talked about individual use-cases of AI in a matter-of-fact way.
They’re also not letting politics get in the way. Committee members approved most of the bills at the hearing by unanimous vote, usually after civil, thoughtful discussions.
Could it be that our state lawmakers are being reasonable for a change? Or maybe AI is so new that the political parties haven’t fully staked out their “die on this hill” positions yet?
Either way, what we’re seeing in the committee hearings is encouraging.
Since we want every Arizonan to understand how the Legislature is dealing with AI, today we’re highlighting the discussions around two bills that caught our eye.
Both bills deal with knowledge work, specifically the legal profession, and how AI tools could offer a cheaper alternative to hiring a lawyer, as well as protect your privacy should the day come when you need to actually hire a lawyer.
Arizona Rep. Teresa Martinez convincing her colleagues to support AI divorces for the poor.
“Divorce for poor people”
Republican Rep. Teresa Martinez introduced HB2371, a bill that would allow people to use AI-assisted arbitration during simple divorce proceedings, as long as they don’t have minor children. A judge would still have the final say.
Martinez said the idea for the bill came from a constituent who told Martinez about how awful her divorce played out.
“He cheated and was not a very good husband and she had decided to leave him,” Martinez said. “It was a clean divorce. They only had vehicles and they only had a house. No minor children. And yet he got everything in the divorce because of the mediation.”
Courts use what’s known as a “child support calculator” to figure out how much the mother or father should pay.
“I thought to myself, ‘What if we had a divorce calculator like we have a child support calculator?’” Martinez said.
The bill wouldn’t cover complicated divorces where the couple is divvying up “money in the Cayman Islands,” Martinez said.
“This is just for simple divorce for poor people and for women and for men who don’t have the means to go out and do a lawyer,” Martinez said.
Democratic Rep. Betty Villegas commiserated with the story of Martinez’s constituent. She told the committee about a friend who got divorced when they were young and the wife had no idea she could have access to his retirement benefits.
Republican Rep. James Taylor, after getting Martinez to commit to fixing what she called a “typo” in the bill-drafting process that made it sound like AI-assisted arbitration would be binding, gave the bill a thumbs-up.
“I think it’s a good idea because it allows people without the means to hire professional people to help in their cases get some intelligent suggestions for what you should expect in the outcome of a divorce,” Taylor said.
Democratic Rep. Junelle Cavero brought up concerns about privacy and Martinez said she’d be happy to add privacy protections into the bill.
Republican Rep. Nick Kupper called the bill a “fantastic start” and said it would be “very worthwhile to get right.”
With that, the committee voted 7-0 to advance the bill.
(Un)common law
To understand a separate bill, HB2410, you have to travel back nearly 1,000 years to the Norman invasion of England.
At least, that’s what the bill’s sponsor, Republican Rep. Alexander Kolodin, seemed to think.
He gave the committee members a history lesson about the emergence of common law after the 1066 conquest as he tried to persuade them to support HB2410.
Arizona Rep. Alexander Kolodin nerds out on the finer points of common law.
The gist of Kolodin’s bill is that when an Arizonan asks an AI chatbot for legal advice, that interaction should be privileged in the same way as conversations with a real lawyer. (Kolodin is a lawyer himself.)
Kolodin grounded his argument in the common law that developed after 1066, what he called “the world’s first self-improving super intelligence” that was able to evolve with each new case that came before a judge.
One of the most important protections that grew out of common law was the protection of a citizen’s privacy, he said. That protection allows people who seek legal or medical advice to know that those communications are protected from discovery if they ever find themselves in litigation.
“The concept is simple. For somebody to get the best medical treatment, somebody to get the best legal advice, they have to be able to have a full and frank conversation and lay out all of the applicable facts, including things that they might have done wrong, maybe things that they didn’t even know that they had done wrong. And that is where the concept of privilege comes from.”
The world obviously has changed quite a bit since the 11th century.
“Now we have teenagers seeking mental health advice, having conversations with AI of the sort they might have once had with a therapist,” he said. “Or maybe they’re discussing a session they had with a therapist with the AI. We have people seeking legal advice from AI, going and asking ChatGPT or Google, ‘Can I do this? Can I do that? Would it be legal if I did this?’”
Without legal protection of their privacy, many of those conversations are an “open book” to anybody who might sue them one day, he said.
What HB2410 does, he said, is to take the recent developments in AI and “port” them into the common law system, which would give courts the permission to start developing common law around the concept of AI privilege.
As Kolodin put it, case law moves a lot faster than the Legislature and speed is a necessity when dealing with the world of AI.
Justin Yentes, who works with the Arizona Attorneys for Criminal Justice, pointed out that lawyers regularly use Adobe AI Assist to develop legal documents. He worried officials could issue a subpoena to an opposing party’s law firm’s data provider and figure out their legal strategy.
Cavero said the bill was “at the forefront” of the expanding area of law that deals with AI, including with privacy.
Republican Rep. Justin Wilmeth, who chairs the committee, said he agreed with Kolodin that the legislative process is slow.
“But now we’re in this age of high-tech, fast-paced stuff and we will be dealing with AI issues next year that do not exist right now,” Wilmeth said. “So we’re always going to be trailing this no matter what we do.”
The bill passed out of committee by a unanimous vote.
Both bills have a few more stops before they reach Gov. Katie Hobbs’ desk and we’ll help you follow along as they move through the Legislature.
As always, if you have a hankering for following the bills yourself, you can use the tracking list we made with our legislative intelligence service, Skywolf.
Preparing for disruption: Another bill discussed at last week’s meeting of the House AI committee would create a statewide AI education program, Tom Blodgett reports for the Scottsdale Independent. Republican Rep. Alexander Kolodin said his bill, HB2409, addresses the “massive disruption to the labor market” that he expects AI to cause “in the very near future” by creating a summer course to teach digital literacy and AI-driven entrepreneurship skills.
Speaking of AI training: Mesa officials are making their city the first in the state to train residents and city employees to use AI, Cecilia Chan reports for the East Valley Tribune. The city is offering free courses through the public library system and city’s LinkedIn Learning account. Several dozen city employees have been taking the courses and next month every city employee will start taking them.
New rules incoming: After the data center project known as Project Blue created heated controversy in Tucson last year, the city is developing new rules for data centers. Tucson officials are hosting public meetings to get feedback on proposed zoning regulations and they’ve put together a Technical Advisory Committee to guide the process. The first meeting is today and you can register here.
Taking it too far: Real estate agents in Arizona are leaning on AI tools to write listings and digitally stage homes, and that’s creating some ethical problems, 12News’ Trisha Hendricks reports. One house hunter found an image of a house that was so badly doctored by AI that it looked like “they basically virtually remodeled the house.”
A digital treasure trove: A team of researchers at Arizona State University are using AI to convert decades of notes written by famed primatologist Jane Goodall into a searchable database, Joycelyn Munoz writes for ASU News. Goodall lived among chimpanzees in Tanzania and ASU’s AI Acceleration team is using machine learning and imaging techniques to extract data points from her notes.
“They help us understand more about human origins, and more about the complex nature of chimpanzee behavior and ecology. A better understanding of their biology and behavior gives a better chance of protecting this iconic endangered species,” researcher Ian Gilby said.