Shaping the Future of AI in Your City

By Suzy Frisch

There is powerful promise in using artificial intelligence (AI) in municipal government operations. AI can help to complete a task like writing frequently asked questions in five instead of 50 minutes, though revising is required. Or start a proposal for a municipal swimming pool with a draft generated by AI, then fine-tune from there. But cities have many details to work out to determine the best and safest ways to use AI — and many are unsure where or how to start.

AI tools already are embedded in our work and personal lives. There’s a high likelihood that employees engage with it at home (Hey, Alexa!) or work, like when text auto-populates in an email. Given this reality, cities should be initiating shortterm policy discussions about artificial intelligence as it exists now, as well as future-oriented conversations about how the tools might affect communities in 10 to 20 years.

At the League of Minnesota Cities 2024 Annual Conference in June, futurist Hauson Le will present a pre-conference workshop about artificial intelligence. Le, strategic foresight manager at Strategic Government Resources (SGR), will engage participants in discussions, debates, and strategy development about AI. The intention is to reveal insights about AI, its risks and ethical implications, and best practices for effective usage.

“We will be learning about the basics of artificial intelligence and holding space to have conversations about big issues like data privacy or the environmental implications of AI,” Le said.

The workshop is rooted in efforts to help municipalities build capacity in innovation and foresight — an approach that emphasizes proactivity and anticipatory government. Such work helps cities understand what is changing and emerging in society and getting ahead of those changes, said Heather Benoit, executive vice president of strategic foresight at SGR.

Starting the conversation

When thinking about artificial intelligence, Melissa Reeder, chief information officer for the League of Minnesota Cities, sees everyday potential with generative AI tools. By using generative AI tools like Copilot or ChatGPT, city workers can jumpstart the draft of a presentation or query risk mitigation strategies for an upcoming municipal project.

“That’s in a nutshell what I call augmenting the city worker,” Reeder said. “AI is not about replacing city workers at all. Instead, it augments them so that they can focus on other value-added activities to serve the public. Generative AI tools do not understand reasoning, emotions, or current events. The output of generative AI is a representation of the massive amounts of data it was given, therefore human interaction is necessary.”

For cities and elected officials that want to consider AI, there are some initial steps to take. Le refers to recommendations from the National Association of State Chief Information Officers (NASCIO) for governments that want to start implementing AI, including:

• Start the conversation about potential pros, cons, and opportunities.
• Establish governance and oversight processes.
• Explore and evaluate what other cities are doing, while considering their city’s community and context.
• Consult with people working on the front lines about potential uses.

There are many evolving factors that will determine how AI and its capabilities develop in the coming years. Significant unknowns include how much computer power, electricity, and water are needed to use AI at a high level.

“Can we get the costs down to make it economically feasible? Outside of that, the sky is the limit,” Benoit said. “It’s up to each local government to choose what flavor of AI they want to implement.”

Local government can focus on employing AI to take advantage of automation or efficiency, or it can use the tools to augment their operations. Benoit points to the possibility of using AI to automate segments of permit review. Instead of attempting to solely speed up the process, cities could explore entirely new ways to approach and redesign permit review to better serve all parties. Other promising areas include using AI for property assessments, enhancing 311 services, and improving search functions to make city information more accessible.

“We have this tendency to think about AI as this cool new tool that can help us get today’s work done faster, cheaper, more efficiently. That’s the wrong way to think about it,” Benoit said. “It really represents the opportunity to completely revolutionize the processes you’re doing and redefine the work you’re doing and do them in a completely different way that is more beneficial or a better experience.”

A holistic review

Many cities are playing policy catch-up because AI is already in use. One resource to consult is the League’s website, where you’ll find information about artificial intelligence and details about what cities should know when using AI at lmc.org/AIconsiderations.

“The genie is out of the bottle, and cities are backpedaling to get policies in place to protect their information,” Reeder said. ‘Rather than a policy to ban AI, cities should consider a ‘yes, and’ policy. Yes, you can use it, and here are the best ways to use it. It’s already here, and it’s important get a policy in place.”

There are some legal and ethical considerations for cities and elected officials to monitor. It’s essential to view AI as a tool, with humans still needing to augment and review its output. That’s integral to avoid bias in data or to prevent nonpublic data getting integrated into AI models, which then can be accessed by other AI systems, Reeder said.

Bias can occur when the data isn’t complete, often because there is missing information about marginalized communities.

“If you’re using a free model on the internet, you might not have a clear vision or accurate results because it doesn’t represent all of the communities you’re trying to represent,” Reeder said. “It’s only as good as the data you collect.”

Other AI concerns include potential breaches of data privacy and workforce displacement as organizations become more efficient. That freed-up time raises additional questions — and potential for significant change.

“At a high level, AI promises to create more capacity, save time, become more efficient, and solve problems,” Le said. “I find myself wondering, if it saves time, does that change where people apply the time they saved?”

Using a foresight planning approach enables cities to think anew about their organizations’ operations. Perhaps new models would remove silos and allow the workforce to rethink how to best use this extra time in new ways, on behalf of all stakeholders, Le said.

“The metaphor is that technology is our savior, but I would say that’s not the whole solution. You still need to have a culture where people are willing to use it,” he adds. “I recently read a study that said removing obstacles to progress is more effective than adding pressure,” Le said. “We’re trying to remove obstacles to progress. Having these conversations can do that.”

Suzy Frisch is a freelance writer.