Like many of you in local government, my inbox now serves as a full-time AI hype repository. Vendors promise the moon, Mars, and a minor galaxy if I’ll just sign the dotted line before lunch. Meanwhile, national headlines swing wildly between utopia and dystopia, surveillance, sentient agents, data centers you can see from space. It’s all rather dramatic.
But from where I sit, especially in rural counties where resources are fewer and patience thinner, the most promising path forward isn’t particularly flashy. In fact, it’s downright mundane. And that’s precisely why it works.
After months of watching public sector AI experiments like a policy wonk with a sociology degree, I’ve sketched out a local leader’s playbook. One that builds capacity without inviting legal headaches, community uproar, or a 7-year proprietary software commitment you’ll regret by next Wednesday.
The Great Divide: What I’m Chasing (and What I’m Politely Avoiding)
At the risk of oversimplifying, I see two very different paths for AI adoption in the public sector.
Path One is the headline-maker: facial recognition, predictive policing, surveillance systems that promise efficiency and deliver controversy. These projects tend to attract lawsuits, protests, and FOIA requests in bulk. They also assume your IT department has unlimited bandwidth and your legal counsel enjoys living dangerously.
Path Two, the one I recommend with the quiet confidence of someone who’s read one too many audit findings, is internal productivity. Yes, the boring stuff.
I’m talking about:
- Drafting and summarizing dense reports (the kind that previously required caffeine and a small prayer),
- Translating documents to actually reach all your constituents,
- Powering internal helpdesk bots to cut down on “Did you try restarting it?” tickets,
- Analyzing zoning codes with fewer migraines.
These aren’t headline-grabbers, but they work. They save time, reduce manual labor, and, here’s the real kicker, they can be implemented without your residents storming the next council meeting with pitchforks.
Start here. Build trust. Demonstrate ROI. Then, and only then, consider anything involving algorithms and citizens in the same sentence.
Why I Treat Governance as an Advantage, Not a Buzzkill
Let’s be honest: policy always lags behind technology. In local government, it sometimes shows up three fiscal years late wearing a “Hello, World” button.
But here’s the contrarian view: having to define our AI standards now, before we go too far, isn’t a nuisance. It’s a strategic advantage.
I’m not proposing a 100-page AI governance manifesto that requires a decoder ring. Just a clear, functional framework that spells out how we’ll handle data, transparency, audit trails, and human accountability.
The message to our communities? “We’re not just deploying shiny tools, we’re thinking first.” In an era where public trust is often in shorter supply than server capacity, that matters.
Retrofitting guardrails after something goes wrong is expensive, in dollars, credibility, and meeting minutes.
Be Vendor-Aware, Not Vendor-Led
We’re all under pressure to do more with less. And vendors know it. Their pitch decks now feature something called “agentic AI,” which, best I can tell, means giving a chatbot your job description and hoping for the best.
But here’s the issue: much of what’s being sold today is still half-baked, opaque, and designed to ensure the vendor, not you, ends up with the most leverage.
The leaders I admire aren’t buying promises, they’re building in-house competence. They’re creating just enough AI fluency to ask hard questions, spot inflated claims, and retain negotiating power.
This doesn’t mean becoming an AI lab. It means being smart enough to stay in the driver’s seat, rather than duct-taped to the hood.
The Rural Reality: Scaling Through Shared Services
Let’s talk capacity. Or more accurately: the complete and utter lack of it.
Rural counties aren’t hiring squads of AI ethicists anytime soon (though I’d personally love to see that job posting). Even if the interest is high, the staffing and budget realities are sobering.
This is why I believe the future lies in shared services. Not just as a nice-to-have, but as the only structurally viable model for much of local government.
I’m working on collaborative approaches like the Digital Innovation Collaborative Exchange (DICE). Think of it as a public-sector co-op for AI, automation, analytics, communications; a place where jurisdictions can pool knowledge, vet tools together, and co-own infrastructure that none of us could afford solo.
Why should 30 towns each reinvent the wheel, and pay retail for it, when we could build the wagon together?
My Bottom Line for Main Street
Here’s the playbook I’m betting on:
- Start small with internal productivity to build credibility and momentum.
- Lead with governance, not as a box-checking exercise but as a public trust asset.
- Steer clear of the backlash zones until you’ve built your own internal competence.
- Collaborate, because going it alone is a good way to waste money, energy, and several perfectly good Tuesday evenings.
The future of public-sector AI isn’t shiny. It’s solid. Shared. Trusted. Less like a gadget and more like a road: invisible when it works, deeply missed when it doesn’t.
And if we get this right? We might just help government feel a little less like a slow-moving bureaucracy and a little more like what it was always meant to be: a system that quietly, intelligently, gets things done.