I don't think I need to convince any public servant when I say one of the biggest challenges of working in government is having to explain and justify decisions. And often against an uninformed public (and sometimes elected officials).
This is especially hard in cases where it's a hot topic on social media. (As a random plug here: if you want to leverage or get ahead of your social media presence, reach out to Sam Toles at CiviSocial. He's great. Note: I have zero financial interest. I just think Sam is stellar.)
AI is quickly becoming one of these topics. Data center conversations are all over the internet. Proponents of data centers argue that foreign influence is fanning flames on social media. Detractors argue over environmental issues in the pursuit of AI. And that's just the beginning.
In the private sector, there are fears of AI being used to displace labor and create mass job loss. Meanwhile, those of us in government have been watching demands rise, budgets get cut, and staff decrease over the past three years. Data shows these trends will worsen over the next five years. So to many in government, AI offers an amazing opportunity at a critical time. Nearly every government we speak with expects to use AI in a significant capacity by 2027.
The worst outcome would be not getting ahead of the messaging with the public and elected officials. And, in turn, having to roll back obvious benefits for your organization. (Imagine saying, "We're going to get rid of all technology and go back to paper." That's how we'll look at this in years to come.)
So how do you get ahead of the narrative?
There have been a lot of headlines recently about Flock Safety. Many are concerned about the surveillance, while others argue for the safety benefits. Regardless of where you land, there is a glaring governance issue underneath it all. Access is too broad, rules about who could see what are unclear, and agencies are discovering (after the fact) how their data was being used.
That's the lesson. Most public blowups about government technology aren't really about the technology. They're about the absence of guardrails. If your community's first exposure to your AI program is a headline, you've already lost the framing. If their first exposure is a clear, published policy—what the AI can do, what it can't, who's accountable, how residents can flag problems—you've set the terms of the conversation before anyone else can. Write the governance before you write the press release.
As we all know, the private sector and public sector are not the same. And this is especially true around the leading narratives around AI. The private sector narrative is dominated by profit at the expense of many. Job loss for higher profits is a common theme.
While in the public sector, the focus has been on expanded and higher-quality service delivery. Government has a completely different opportunity with AI, and it's important that agencies highlight that this narrative is different in their communications.
I guarantee not a single agency would say "we get to do everything we want to do." Every clerk, every planner, every public works director has a backlog they'll never clear at current staffing. If you could save 50% of your time on the phone calls, the records requests, the repetitive questions, that time doesn't turn into layoffs. It turns into the work that's been sitting undone for years. And there's a benefit a lot of organizations don't expect: retention. When people get to spend their time on work that actually matters, they stick around.
In the private sector, efficiency flows to shareholders. In government, it flows back to residents. That is the story, and nobody will tell it for you. When you talk about AI publicly, don't lead with "cost savings." Lead with "the permit that used to take three weeks now takes three days. And not only did nobody lose their job, the team now spends more time guiding people through what they want to build instead of reading through paperwork to reject things that don't meet code."
You don't have to figure this out alone, and honestly, you shouldn't. Your peer agencies, your state municipal league, your county association: Someone near you has already had the town hall meeting you're dreading, already answered the angry public comment, already written the policy you're staring at a blank page trying to draft. Borrow shamelessly.
The same goes for vendor partners. Any partner worth working with should be able to hand you their security documentation, explain in plain English where your data goes, and, if asked, stand next to you in front of your council. If a potential partner gets cagey when you ask them to help you explain the technology to the public, that tells you something. The partners who are good at this want to be part of your public conversation, because their success depends on your community trusting the program.
So if they can't deliver on all that, they’re likely not a partner. They are just a vendor.
Most fear of AI is really fear of the unknown, and the unknown is fixable. Hold the demo. Invite your council to watch a resident ask the assistant a question and get an answer at 11 PM on a Sunday. Show your staff what the tool actually does. Not just a slide about it, show the actual thing. I've watched skeptical elected officials completely change their posture in 15 minutes, not because someone won an argument, but because they saw a resident get help.
And educate continuously, not once. Publish what the AI handled last month. Share where it fell short and what you fixed. The agencies that treat education as an ongoing habit rather than a launch event are the ones that never end up on the defensive.
Every public servant knows this dynamic: three people at public comment can sound like the whole town, especially when they're amplified online. But the resident who got their question answered at midnight, or renewed a license without taking a morning off work, doesn't come to the podium. They just quietly have a better experience.
Your job is to make the silent majority visible. Collect the numbers: how many questions answered after hours, how many calls deflected, how much faster the response. Collect the stories, with permission, and tell them. When a critic says "nobody wants this," the best answer isn't a rebuttal. It's "4,000 of your neighbors used it last quarter." Take the concerns seriously, fix the legitimate ones, but never mistake volume for consensus.
Finally, there is no universally right first project, but there's almost always a right first project for your community. If your community is privacy-sensitive, don't start anywhere near public safety. Start with something boring and obviously helpful, like answering questions about trash pickup and park hours. If your community's biggest frustration is that nobody answers the phone, start there, because the win will be felt immediately by the people most likely to complain.
The best first use case is one where the benefit is obvious, the risk is low, and the story tells itself. Get that win, tell that story, and the second and third projects get dramatically easier because now you're not asking your community to trust a hypothetical. You're asking them to let you do more of the thing that's already working.
The public will use AI agents to interface with government. You can't afford to wait.
Here's the part I think people underestimate. This conversation isn't just about whether government adopts AI. Your residents already have. They're asking ChatGPT how to appeal their property assessment and what your zoning code allows. Soon their AI agents will be calling your phone lines and filling out your forms on their behalf, whether you've prepared for it or not. And remember: these agents are private sector products. You don't control them, you can't configure them, and you don't get a say in what they tell your residents.
That flips the question. It's no longer "should we adopt AI?" It's "when an AI reads our website or calls our office, does it get the right answer?" If your information is buried in PDFs from 2019, the AI your resident is using will confidently give them the wrong answer, and they'll blame you, not the chatbot. Getting your own house in order isn't just an efficiency play anymore. It's how you stay the authoritative source for your own community.
In conclusion, the agencies that navigate this well over the next few years won't be the ones with the best technology. They'll be the ones who treated the public conversation as part of the work and not an obstacle to it.
Get ahead of the narrative, and the narrative works for you.
.avif)
