‘Artificial intelligence is not innovation,’ it’s a tool. How governments use it will vary.
Nutthaseth Vanchaichana/GettyImages
Some state and local governments may embrace AI wholly, while others may take a more measured approach. Either way, experts said, the competition to be first is moot.
Announcing its ambitions to be a “global hub” for artificial intelligence, California was the first state to regulate its use by state agencies. Boston was one of the first cities to issue guidelines for how its workers can use generative AI. And Tempe, Arizona, turned heads last June when it enacted what is believed to be the first policy on AI’s ethical use.
Recognizing its economic and cost-savings potential, everybody wants to be first to harness AI. But while the race to embrace the new technology may leave some government leaders feeling like they are already lagging behind, they needn’t worry so much, according to David Graham, chief innovation officer for the city of Carlsbad, California, and co-chair of the Civic Innovation Executive Certificate program at the Technology and Entrepreneurship Center at Harvard University.
Graham said it’s still too early in policy development for governments to measure themselves against each other. What’s more, every government’s AI needs and uses will be different, making any sense of competition moot.
“If you don't have a generative AI policy right now, you're not missing out,” Graham said at a conference hosted by the International City/County Management Association, or ICMA, this week in Boston. “Artificial intelligence is not innovation. Technology is a tool. AI is a tool. Generative AI is a tool. As soon as we begin to think of it in that particular way, then it's much more about the application of it in our context, than it is about just using the thing because that's what everybody else is doing.”
Many cities are currently experimenting with various use cases for AI, including automating repetitive tasks, powering call centers and chatbots. Noting that it is tempting for governments to share best practices with each other, especially on use cases for new technologies, Graham cautions against that. His reasoning? Each city’s circumstances are so different, especially the level of investment it can make, how large its operations are and the impact the tools could have.
“Oftentimes in technology or economic development or in these other things, it's like, ‘Oh, well, those cities did it, where this is the best practice or this is what we shouldn't be doing,’” he said. “And we feel like we're having to stretch and reach. This is the space, I would say, [where] that is not going to be the case.”
The race to AI is not like competition in other areas, Graham continued. Instead, it is a “spectrum” of different deployment levels, comfort, technology maturity and standards.
Take public benefits and public health programs, which could improve with help from AI to process applications more quickly and get money out to residents. Especially for larger agencies, the technology could help alleviate backlogs. But smaller agencies that are cash-strapped, especially in rural areas, may be best served to deploy the technology and its limited money elsewhere.
“We can't think of the United States as this one level playing field [where] everybody has access to all these wonderful tools and everybody listens to them the same way, follows the advice the same,” Lori Freeman, CEO of the National Association of City and County Health Officials, said during an ICMA panel discussion. “It's very different from jurisdiction to jurisdiction across this country. We must be careful that we don’t unilaterally layer a tool on our health departments, at least that's if not all of them could take advantage of or utilize the same way.”
Graham said local governments should not wait for the federal government, or even states, to act on AI. They must do it themselves.
“This for me comes down to the same thing that we usually find at the local government level,” he said. “It's up to us, no one's coming to save us, we’ve got to figure it out ourselves. So collectively, we are figuring it out.”
NEXT STORY: Election laws fighting AI deepfakes need to be targeted and adaptable, report says