Lawmakers may soon begin drafting legislation aimed at implementing or restricting the use of artificial intelligence after leaving the National Conference of State Legislatures’ annual summit.

The burgeoning technology was a major focus of this year’s conference, which featured nearly a dozen AI-focused sessions. Lawmakers from statehouses across the country learned about cybersecurity, election interference, workforce development and other topics.

After one session, an informal poll showed 87% of attendees felt their state laws were not adequate to deter the threat AI could pose to elections. 

NCSL Associate Director Ben Williams said in an interview that there have been hundreds of AI-related bills introduced in recent years — and lawmakers are hungry to learn more. 

“This is something legislators know is a really important issue going into the coming years, and it’s not going to go away,” he said. “It’s only going to grow in importance, and so they told us to try to put more information up about AI, and we’re trying to deliver as best we can.” 

The conference’s focus was clear from the beginning when the NCSL booked Jennifer Pahlka, the founder of Code for America, as the opening speaker on Day 1. 

Pahlka suggested thinking about AI as if it’s an intern. 

“You may have some fantastic interns, but you’re always going to have a more senior staff member look over anything that your interns might write and make sure that it is correct and make sure that it is consistent with you and your values and what you’re trying to do for the people that you serve,” she said. “And I think AI, especially generative AI, can be a great tool for all the things that you’re doing.”

AI to lessen burden, optimize results

There are three ways to improve state capacity, Pahlka told lawmakers and other attendees: focus on the right things, have the right people and burden them less. 

“If you can use AI for simplification, you can burden folks less to do more,” she said. 

Pointing out the length and complexity of unemployment insurance codes, Pahlka said, “AI is going to help us make sense of that complexity. 

Connecticut state Sen. James Maroney, who’s led the way on AI issues in his state, said there’s a rush to implement the new technology.

It’s a “shiny new thing,” he said, but it’s important to recognize states are “very different from the private sector” because governments don’t have the same ability to take risks.

Maroney pointed to scenarios where government AI use went wrong, including when Spanish law enforcement used the technology to assess the risk of domestic violence, leading to a woman’s death at the hands of her husband after he was rated as low-risk.

Alexi Madon, a government relations executive at IBM, and Jamia McDonald, a principal in Deloitte’s Human Services Transformation division, both said states are already using AI in many areas — even if they don’t realize it.

States should start by inventorying and disclosing their government’s existing uses of AI, Maroney said, a policy his state has already implemented.

McDonald and Madon, who joined Maroney on a panel on AI’s use in government, said it’s crucial for states to adopt a risk management framework to determine what tasks they are and aren’t comfortable with using the technology to accomplish.

“When you have a security framework, it is easy to get to the right outcome,” McDonald said.

Madon suggested agencies hire their own AI-specific experts instead of relying solely on existing IT staff.

Is artificial intelligence close to a point at which it could be used to draft bills?

“Do I think it’s happening tomorrow? No,” Madon said. But she said that possibility isn’t too far away.

AI in the Wild West of politics

While some sessions highlighted the ways AI can improve government, others focused on how it can be used to sow chaos.

“That’s a wild future that we’re looking at,” Kentucky state Sen. Whitney Westerfield said about the possibility of AI-written bills.

New Hampshire Secretary of State David Scanlan knows those downsides all too well after a political consultant sent AI-generated robocalls that sounded like President Joe Biden dissuading voters from casting votes 24 hours before the statewide primary. 

Very quickly after it was reported, the state’s attorney general made a public statement that the phone call was misinformation and a “severe form of voter suppression.” 

“My office reinforced that message, and we kept delivering that message throughout the day,” Scanlan said.

He considers the response a success, saying, “If there was an impact, it was not on a scale that affected the outcome of the election.” 

New Hampshire now has a law requiring a disclaimer on AI-generated messaging.

Cait Conley, senior adviser to the director of the U.S. Cybersecurity and Infrastructure Security Agency, said AI has positive uses.

“[But] we can’t blindly accept the good without recognizing the potential for bad,” she said. “Just like every other technology, there’s going to be malicious actors who want to further their own objectives.”

The technology can also be used by foreign adversaries — such as Russia, China and Iran — to foment partisan distrust through misinformation campaigns that have migrated from public platforms to chat channels, she said.

Conley recommended states ensure constituents know where to find reliable information and said one easy step is to transition all local election board websites to a  “.gov” domain, which has been free since 2022.

The potential consequences of AI are real, McDonald said. She cautioned that although it’s a new and exciting tool, lawmakers should first determine exactly what problem they’re trying to solve — and whether AI is the right solution.

“Just because you can, doesn’t mean you should,” she said.

Krista Kano is a staff writer for Gongwer Ohio/State Affairs. Reach her at [email protected] or on X @krista_kano. 

Brett Stover is a Statehouse reporter at State Affairs Pro Kansas/Hawver’s Capitol Report. Reach him at [email protected] or on X @BrettStoverKS.