AI for Public Good: How Cities Balance Innovation and Ethics

AI in Municipal Operations: Practical Use Cases Driving Change
Cities across the globe are applying AI to improve core services, with transportation, public safety, and resource management leading the list. In transportation, AI-powered traffic signal optimization systems are easing congestion and reducing emissions. For example, Pittsburgh deployed adaptive traffic signals through the Surtrac system, which uses real-time data and AI algorithms to adjust light timings dynamically. This initiative reduced travel times by over 25 percent and idling time by more than 40 percent, demonstrating how AI can improve both efficiency and environmental outcomes^1.
Public safety is another area where municipal leaders are leveraging AI tools. Predictive policing platforms, such as those trialed in Los Angeles and Chicago, analyze historical crime data to allocate patrols more effectively. However, these systems have raised ethical concerns, especially around disproportionately targeting communities of color^2. Cities like Seattle have taken a different approach by using AI to monitor gunshot detection systems and emergency call data to improve response coordination without relying on predictive profiling^3. These applications show AI’s potential to support, not replace, human decision-making in critical safety contexts.
Ethical Considerations: Data, Bias, and Consent
Implementing AI in city governance introduces significant ethical challenges that must be addressed from the outset. One core issue is data privacy. AI systems depend on vast quantities of data, often collected from sensors, mobile devices, and digital interactions. Without clear data governance policies, cities risk eroding public trust. For instance, Toronto’s Sidewalk Labs project faced public backlash due to concerns over surveillance and unclear data ownership, ultimately leading to its cancellation^4. This case illustrates how lack of transparency around data use can derail even well-intentioned smart city projects.
Another pressing challenge is algorithmic bias. AI systems trained on historical data can perpetuate existing inequalities unless explicitly corrected. For example, facial recognition software used in public surveillance has shown significantly higher error rates for people of color^5. In response, San Francisco became one of the first U.S. cities to ban the use of facial recognition by government agencies, citing concerns over accuracy and racial profiling^6. These policy decisions reflect a growing awareness that ethical AI is not only about technical performance but about aligning technology use with community values.
Models of Responsible AI Governance in Cities
Cities that are leading in responsible AI adoption have implemented governance structures that prioritize transparency, community engagement, and oversight. One effective model is the use of citizen advisory boards. In Helsinki, the city has established an AI Register that publicly lists all municipal AI systems and explains their purpose, data sources, and logic in plain language. This initiative is complemented by community feedback channels that allow residents to question and challenge algorithmic decisions^7. By involving the public directly, Helsinki has fostered a higher degree of trust and accountability.
Another approach is adopting transparent data-sharing frameworks. Amsterdam has developed a system that tags all AI applications with “Algorithm Registers,” similar to nutrition labels, that disclose what data is used, how decisions are made, and who is responsible for oversight^8. These registers are accessible online and updated regularly. This model empowers residents and civil society groups to monitor the ethical behavior of automated systems. By embedding transparency into the design of AI tools, cities can proactively mitigate risks while maintaining public confidence in innovation.
Actionable Recommendations for Municipal Policymakers
To harness AI responsibly, municipal leaders must establish clear frameworks that balance innovation with accountability. First, cities should adopt AI ethics guidelines grounded in human rights and equity principles. These should be developed collaboratively with residents, technologists, and legal experts. The City of Montreal, for instance, has published an AI ethics policy that guides all municipal deployments and includes public consultations as a core requirement^9. Embedding ethics from the start helps prevent reactive policy adjustments later.
Second, policymakers should build internal capacity by training staff on AI literacy and ethical risk assessment. Many departments lack the technical background to evaluate AI vendors or manage data risks. Establishing cross-functional AI review committees, with representation from IT, legal, equity, and community relations teams, can help ensure that decisions are not made in isolation. Finally, cities should require vendors to meet procurement standards that include transparency clauses, bias testing, and post-deployment audits. These measures ensure that AI systems remain accountable throughout their lifecycle.
Building AI Systems that Earn Community Trust
As cities continue integrating AI into everyday services, the path forward must be guided by the principles of trust, transparency, and inclusion. Trust is earned when residents feel their data is protected and their voices are heard. Transparency means more than publishing technical documentation—it requires clear communication, accessible explanations, and open channels for feedback. Inclusion demands that AI systems serve all communities equitably, particularly those historically marginalized in civic processes.
Municipal leaders are uniquely positioned to shape the future of AI through values-driven governance. By centering human needs, embracing accountability, and involving the public in decision-making, cities can align technological advancement with democratic ideals. AI is not a replacement for civic leadership—it is a tool that, when used responsibly, can help leaders build cities that think with empathy and act with integrity.
Bibliography
Smith, Stephen. "Real-Time Adaptive Traffic Signal Control System: Evaluation of Surtrac." Carnegie Mellon University Traffic21 Institute, 2019.
Ferguson, Andrew Guthrie. "The Rise of Big Data Policing: Surveillance, Race, and the Future of Law Enforcement." NYU Press, 2017.
City of Seattle. "Real-Time Crime Center Operations and Data Integration Strategy." Seattle Police Department Strategic Plan, 2021.
Goodman, E. P., and Powles, Julia. "Urbanism Under Google: Lessons from Sidewalk Toronto." Fordham Law Review 88, no. 2 (2019): 457-498.
Buolamwini, Joy, and Gebru, Timnit. "Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification." Proceedings of Machine Learning Research 81 (2018): 1-15.
San Francisco Board of Supervisors. "Stop Secret Surveillance Ordinance." City and County of San Francisco, 2019.
City of Helsinki. "AI Register." City of Helsinki Digitalization Unit, 2023. https://ai.hel.fi
City of Amsterdam. "Algorithm Register." Amsterdam Data Lab, 2023. https://algoritmeregister.amsterdam.nl
City of Montreal. "Declaration of Responsible AI in the City." Montreal Office of Smart and Digital City, 2020.





Join the discussion
Members can read and add comments on every article.