Understanding Municipal AI Governance: Practitioner Perspectives on AI Policy and Enactment
Drawing on grounded theory and interviews with practitioners across five regions, this study reconceptualizes municipal AI governance as a cyclical, adaptive process shaped by practitioner judgment and institutional context, identifying five interrelated dimensions that explain how policy meaning is constructed in everyday practice rather than merely implemented as static rules.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
The Big Picture: The "Street-Level" Reality of AI
Imagine a city government as a giant, busy kitchen. For a long time, experts have been writing the "recipe books" (policies) for how to cook with new, high-tech ingredients called Artificial Intelligence (AI). These recipe books tell the chefs exactly what to do to ensure the food is safe, ethical, and delicious.
However, this study argues that reading the recipe book isn't the same as cooking the meal.
The researchers interviewed 14 "chefs" (local government workers) from places like Australia, the US, Spain, Saudi Arabia, and Hong Kong. They wanted to know: What actually happens in the kitchen when you try to follow these AI rules?
They found that governing AI isn't a straight line from "Rule Written" to "Rule Followed." Instead, it's more like juggling while riding a unicycle. The people on the ground have to constantly interpret, adjust, and negotiate the rules based on their specific kitchen, their available ingredients, and the hungry customers waiting outside.
The Five Key Ingredients (The 5 Dimensions)
The researchers discovered that municipal AI governance is shaped by five interconnected parts. Think of these as the five stations in the kitchen:
1. Policy Governance: The "Rulebook Source"
This is where the rules come from. The study found four ways cities get their rules:
- The National Menu: Some cities (like in Saudi Arabia) just follow the big national government's menu. They don't write their own; they just execute what the head chef in the capital says.
- The Contract Menu: Some cities hire outside companies to cook. The rules are written into the contract with the vendor. If the vendor breaks the rules, the contract handles it.
- The Hybrid Menu: This is a mix. The city takes the national rules but adds its own "secret sauce" or local flavor to fit their specific needs.
- The Local Menu: Some cities (like Tempe, USA) are pioneers. They write their own rules from scratch because the national government hasn't given them a menu yet.
The Takeaway: The source of the rules changes how much freedom the local chefs have. If the rules come from far away, they might feel rigid. If they write their own, they have more freedom but also more work.
2. Policy Objectives: The "Why We Are Cooking"
What is the goal? The chefs said the rules aren't just about following orders; they serve five main purposes:
- Accountability: Making sure a human is always the one responsible for the final dish, even if a robot helped chop the vegetables.
- Guardrails: Putting up fences to keep the AI from wandering into dangerous territory (like privacy violations).
- Equilibrium: Finding the "Goldilocks" zone—not so strict that you can't innovate, but not so loose that you burn the house down.
- Awareness: Teaching everyone in the kitchen what the new tools are and how to use them safely.
- Enablement: Using the rules to actually help the kitchen run faster and better, not just to stop things.
3. Policy Implementation: The "How We Cook"
This is the actual work of putting the rules into action. The study found five main methods:
- Communication: Constantly talking to everyone to make sure they know the rules.
- Transparency: Keeping the kitchen doors open so the public can see how the food is made (unless there's a good reason not to).
- Compliance: Checking boxes and signing forms to prove you followed the recipe.
- Adaptability: Being ready to change the recipe tomorrow if the ingredients change.
- Training: Teaching the staff how to use the new tools without getting burned.
4. Policy Challenges: The "Kitchen Nightmares"
Even with a good plan, things go wrong. The chefs identified five big headaches:
- Innovation Trade-offs: The fear that if you put up too many safety fences, you stop creating new, amazing dishes.
- Administrative Complexity: The paperwork is so heavy it slows down the cooking. It's like trying to cook a gourmet meal while filling out tax forms.
- Resource Limitations: Not enough money, not enough skilled chefs, and not enough time. Small towns often can't afford the fancy equipment big cities have.
- Stakeholder Unawareness: The staff (and sometimes the public) doesn't understand what AI is or why the rules exist.
- Digital Lag: Some cities are running on high-speed internet, while their neighbors are still using dial-up. This creates a gap where some cities can use AI safely, and others can't.
5. Policy Requirements: The "Shopping List"
To fix the nightmares, the chefs said they need specific things:
- Resources: More money and more experts.
- Ethical Accountability: A clear promise to the public that they will be honest and fair.
- Technical Frameworks: Clear, step-by-step guides on how to handle risky situations.
- Stakeholder Engagement: Talking to everyone involved (staff, citizens, experts) to get their input.
- Institutional Coordination: Making sure the legal team, the IT team, and the finance team are all on the same page.
The Big Discovery: It's a Cycle, Not a Line
The most important thing this study found is that governance is a loop, not a straight line.
Imagine a cycle:
- You start with a Governance Structure (The Rulebook).
- You set Objectives (The Goal).
- You try to Implement it (The Cooking).
- You hit Challenges (The Burnt Toast).
- You identify Requirements (The Fix).
- Those requirements change how you view the Governance Structure, and the cycle starts again.
The people on the ground aren't just robots following instructions. They are sense-makers. They look at a rule, think about their specific city's budget and culture, and figure out how to make it work. If the rule is too hard, they adapt it. If it's too vague, they fill in the gaps.
Why This Matters
The paper concludes that we can't just write a perfect policy and hope it works. We have to understand that the people actually doing the work are the ones who make the policy real.
If policymakers want AI to work well in cities, they need to:
- Stop treating rules as static laws and start treating them as flexible guides.
- Give smaller cities the same resources and support as big cities so they aren't left behind.
- Listen to the "chefs" in the kitchen, because they know where the recipe is missing a step.
In short: AI governance isn't about the rulebook; it's about the people holding the pen and the spoon.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.