Cities Join the AI Regulatory Game: How Local Governance Reshapes the Future of Artificial Intelligence

Cities Join the AI Regulation Game: How Local Governance is Reshaping the Future of Artificial Intelligence

While federal and state governments debate AI regulation, a more grounded governance force is rising—cities. Last week, Rockville, Maryland became the first city in the state to ban the use of algorithms to set rents, following in the footsteps of San Francisco and New York, which have already imposed restrictions on AI rental pricing and hiring decision tools, respectively. According to the Center for Digital Democracy, more than 20 cities and counties across the U.S. have enacted their own AI policies.

This phenomenon reveals a fundamental shift: The application of artificial intelligence is not an abstract technological issue but a concrete practice deeply embedded in the daily operations of cities. From housing rents to public transit, from public safety to employment discrimination, AI algorithms are rewriting the underlying rules of urban life. And local governments are precisely the most direct enforcers of and those affected by these rules.

Why Must Cities Get Involved?

“Cities have always had jurisdiction over certain policy areas. When internet companies enter these areas, city intervention is essentially an extension of traditional regulation,” notes David Schleicher, professor of urban law at Yale University. Just as cities once imposed regulations on ride-hailing platforms like Uber, they are now regulating the use of AI in local, specific contexts.

Stefaan Verhulst, co-founder of the NYU Governance Lab, adds that cities have more extensive authority to oversee “how AI is used in local service delivery, or in areas such as employment and education that have been delegated to local governance.” For example, a city where a self-driving car company operates may not be able to dictate how models are developed, but it can ban the vehicles from running on local roads.

The advantage of this close-to-the-ground governance is that cities are more keenly aware of the impact of technology on real communities. When data centers provide computing power for nationwide AI systems, the resulting energy consumption, noise, and environmental costs are borne by surrounding residents. Aaron Saiger, director of the Center for Urban Law at Fordham University, says: “AI is a phenomenon where the beneficiaries and those bearing the costs often do not live in the same place.” City regulation forces us to confront this geographical inequity.

The Dilemma of Fragmentation: A Trade-off Between Innovation and Democracy

However, the proliferation of city-level AI regulation also brings significant challenges. The AI industry complains that companies already have to navigate a patchwork of state laws, and city-level rules will further increase compliance complexity. More critically, many cities lack staff with technical expertise, making it difficult to formulate truly effective regulatory policies.

But restricting cities’ AI regulatory power also poses problems. Verhulst emphasizes: “Cities can implement more participatory AI governance—they can genuinely consult with communities and residents to understand their expectations for AI use. While states can also do this, cities are closer to residents than states or even the federal government.” This kind of democratic input is particularly important in the AI field, as the risks and benefits of technology are often unevenly distributed across different communities.### From Rules to Participation: The Shift in Future Urban AI Governance

The debate over urban AI regulation is not simply about whether to regulate, but rather how to regulate and who has the authority to do so. Schleicher points out that U.S. state constitutions grant different specific powers to local governments, and most current city charters do not explicitly grant authority over AI regulation, yet they cover traditional areas such as housing and transportation. This means that cities' AI policies are likely to be based on extended interpretations of their existing jurisdictions.

A more fundamental perspective is that cities should be allowed to serve as experimental grounds for AI governance. Just as some cities have pioneered data transparency and algorithm transparency requirements, they have provided valuable practical experience for higher-level policymaking. For example, New York City requires bias audits of AI tools used in hiring decisions, and this policy has become a reference model for other regions.

Meanwhile, the European Union has just announced exemptions from removable battery requirements for wearable devices such as smart glasses, clearing the way for companies like Meta to enter the European market. This development shows that regulatory games at the international level also affect the availability of technologies at the city level.

DeepMind CEO Demis Hassabis recently called on the United States to establish an industry-funded standards body similar to the Financial Industry Regulatory Authority to test and certify cutting-edge AI models. Such top-level design is certainly important, but the practices of cities show that only when regulation reaches the last mile of technology deployment can communities truly benefit or be protected from harm.

Cities and AI: An Unavoidable New Frontier of Governance

Cities joining the debate on AI regulation is essentially an inevitable step in the process of democratizing technology. When algorithms begin to determine who can rent an apartment, who can get a job, and who is liable in autonomous driving accidents, these issues are no longer abstract ethical discussions but require specific, geographically and institutionally grounded responses.

In the future, we may see more cities exploring "participatory AI governance"—through community consultations, public hearings, and transparency audits, enabling residents to directly participate in the formulation of AI rules. At the same time, policy collaboration and information sharing among cities will become crucial to reduce friction costs caused by fragmentation.

Cities are not bystanders in AI technology. They are both the physical spaces where AI is deployed and the smallest viable units for technological governance. When federal and state levels are deadlocked, cities are proving through action that the future of AI can be rewritten starting from the neighborhood.