Prioritizing Vision Over Efficiency Will Make or Break Successful AI Use in Local Government

Micah Gaudet, deputy city manager and public safety officer for Maricopa, Arizona, predicted that like the invention of Microsoft Word, AI will change the way cities conduct hiring and do business.

“[Today] You could have the absolute best person for that job … the right culture fit; they’ve got a good work ethic; they’ve been working in the industry for a while; everything lines up perfectly. If they don’t know Microsoft Word, you’re not going to hire them, and in fact, you’re not even going to interview them,” Gaudet told the audience of more than 400 elected leaders, city staff and residents at the South Bay Cities Council of Governments’ 24th General Assembly. “That’s the reality and has been the reality for the past 10 to 20 years
now in local government … AI is going to have a similar end result.”

Each year, the General Assembly brings thought leaders together to discuss important regional issues. This year’s event was themed Artificial Intelligence: Friend or Foe for the South Bay? Gaudet, who founded the AI and Local Government Network, which connects 200 members from 190 local governments worldwide, was among six speakers to explore how AI could potentially benefit local governments, as well as potential red flags that cities should consider when adopting and using the technology.

In his talk titled “Steering Creativity and Efficiency: Driver’s Education for Generative AI,” Gaudet asserted that when integrating AI into city work, efficiency for efficiency’s sake isn’t always desirable. Instead, he encourages local government professionals to employ AI technologies to be efficient with the “right things.”

As an example of “wrong” things, he recapped the tale of film company Kodak’s demise. While Kodak invented digital camera technology, it failed to see where the industry was going and focused on efficiency over vision, putting “all their chips in physical media.” This led to the company’s eventual

“AI is an accelerator. An accelerator is very, very helpful when you’re moving in the right direction on the highway. If you’re going the wrong direction, it’s downright deadly,” he said. To avoid steering off course, he recommends that organizations stay crystal clear with mission alignment.

During the audience Q and A session following Gaudet’s talk, audience members expressed concerns that use of AI could result in a scaled-down workforce as the technology streamlines work previously done
by humans.

On the contrary, Gaudet predicted that “AI won’t replace people, but people with AI [skills] will [replace them].” He doubled down against losing sight of the mission and prioritizing jobs over creativity and efficiency.

“Our mission is that we are here to do our job. Is our job to save jobs? Or is our job to deliver exceptional public sector services and goods to our residents in the most responsible manner?” he said. “When I wake up in the morning and go to work, I’m not going to work to try and save people’s jobs. I’m going to work to try and make the city the best it can possibly be, and I want people to come on that journey along with me.”


Gaudet advised that the key to successfully navigating the transition to AI, is getting everyone to “row in the same direction.” Like the concept of drivers’ training, alluded to in his talk’s title, he said the goal is to build
confident—but not arrogant—drivers who can navigate different hazards to get from point A
to point B.

“First learn stand; then learn fly,” he said, borrowing a line used in the movie “The Karate Kid” by Mr. Miyagi as he trains a young boy to be a karate master. By this, he suggests that workers start small by slowly exploring AI applications to build confidence. Examples include experimenting with ChatGPT—a chatbot developed by OpenAI that uses large language models to enable users to refine and steer a conversation toward a desired length, format, style, level of detail and language.

Examples of entry tools include using AI to create notes following a virtual meeting. Other simple uses include employing ChatGPT to create a job description or new city policy. He suggests workers use their own preexisting domains of expertise so it’s easy to determine if the technology is providing right or wrong
answers. •

View the 24th General Assembly replay at