So many cities, towns, and counties have a strong desire to move beyond the status quo and embrace opportunities for AI to make their services more efficient, to connect with residents more easily, and to work toward resolving ongoing wicked problems.

But the narrative around AI is sometimes laced with confusion, hesitation, and negativity. When we talk about dispelling myths of AI, the common thread is fear.

“Dispelling Myths and Optimizing Outcomes of AI Solutions” was an education session at ICMA’s Local Government Reimagined (LGR) Conference in Boston. Moderated by Lean Geraghty (founder of Data Spark Consulting and former director of sustainability and innovation at National League of Cities), the session addressed those fears and reframed them as opportunities for local government.

The presenters kicked off the session by establishing a common definition of what we mean when we talk about AI. Melanie McDonough, chief innovation officer of Lebanon, New Hampshire, said, “It’s an assistant that’s there to help guide you along the path of the work you already do — to enhance your work, to fill in gaps, and to help you be more efficient.” It’s one of many tools that we can use in local government to help us do our jobs.

Brianna Sunryd, who leads the Public Innovation Design Studio at the Collins Center for Public Management at the University of Massachusetts Boston, added, “AI is a support player, a collaborator, something that can generate a first draft to get you closer to the finished product.”

Melanie and Brianna highlighted the top myths they hear on a regular basis and how to talk through those myths with their organization and addressing the fear head on.

Unfounded fear comes from a lack of education or knowledge in an area, and training is your biggest way to combat that fear.

Myth #1: AI Is a New Thing.

Or rather, a new thing to most people. It’s been around longer than we realize; it’s just now more in the mainstream and grew in popularity very quickly. We’ve been using forms of AI for years (e.g., virtual assistants like Siri and Alexa, predictive text in emails and texts, spam filters). It’s important to remind your AI-reticent employees of these examples. When we talk about it being new, it’s really Generative AI that’s new, and it’s our responsibility to inform our staff of its potential.

Melanie said, “As leaders, we should never make decisions from a place of fear. Leaders seek to understand. Once we understand it, we need to decide what’s right for our organization and set parameters or procedures. We train our staff and then after some time, we stop and review: How is our work being enhanced through AI? Is it helpful? Has anything come up that we need to address?”

Myth #2: AI Will Replace Local Government Jobs.

“More than anything, AI is a collaborator,” said Brianna. A better way of saying it is that AI will impact local government jobs, not eliminate them. She emphasized that when talking with staff, it’s important to reiterate the value and unique experience that people bring to the organization, like their understanding of the community or being able to adapt in emergency situations. “It all goes back to public trust. Your residents trust the people in city hall.” The distinctly human elements of being in the public service field and interacting with the public are what make our jobs safe.

When Melanie’s town decided they would move forward with establishing an AI policy, she wanted to conduct a survey to see how employees were feeling about AI and what issues the policy should address. In the survey results, a comment that came up over and over was “AI won’t be able to do [this part of my job, that part of my job].” She said, “You need to tell employees that AI is not coming for their job; it’s going to enhance their job and make them a lot better at their job.” Specific tasks may change over time, but the human element will not. In fact, AI often allows staff to focus on the parts of their job they enjoy most.

Myth #3: AI Is Too Big of a Security Risk for Us to Use.

Many communities say that because they don’t have AI expertise in house, they’re worried about cybersecurity. The presenters were quick to say that yes, AI can pose security risks, but so do all the other technology tools we use. With emerging technologies, there is always inherent risk, but with careful consideration, you can go about it strategically. Go back to the policies that your organization already has in place — for social media, email use, data privacy — and start there to craft your guidelines for AI implementation and usage. Consider limiting staff to certain tools, certain instances in which to use AI, or times in which department head approval is needed.

Make a list of your concerns and partner with your IT/cyber services staff. Consider establishing a technology review committee to evaluate AI tools that would be a good fit for your organization. And don’t be afraid to bring in outside help if needed. Then set your policies and train your staff. The presenters agreed that it all goes back to education. Training your employees is the best way to prevent most cybersecurity issues, and it’s the same with AI.

Generally, a good rule of thumb when it comes to sensitive data: If you wouldn’t put it in an email or a message on Teams or Slack, don’t put it in Generative AI. Municipalities are already getting right-to-know requests for ChatGPT prompts. That’s why it’s important for your AI tools to be organization-wide and to not allow personal accounts.

Myth #4: AI is Approaching the Same Level of Human Intelligence.

Many folks incorrectly assume that AI never gets anything wrong. But as Melanie said, “Once you start using generative AI, your first thought may be ‘this thing is so dumb.’” AI output still needs to be double checked for accuracy. AI hallucinations — incorrect or misleading information — happen all the time. A robot with capabilities closely rivaling that of a human is artificial general intelligence (AGI), which most researchers believe is decades if not centuries away. Where we are now is considered “weak AI.”

Ultimately, what is incumbent upon us now is deciding how best to use the AI tools available to us in this day and age to better serve our communities.

Learn more about AI for local government on ICMA's Generative AI resource page or by attending the Local Government Reimagined Conference, June 5–7, 2024, in Palm Desert, California.