“What happens when a city asks ‘Should we’ before ‘Can we?’ with AI?”
In an era where artificial intelligence (AI) can bring unprecedented efficiency to cities, the real challenge isn’t just about how fast governments can adopt new tools but whether they do so in ways that truly benefit their communities. Efficiency is important, but ethical considerations are just as critical, especially when public trust and social impact are at stake. Beginning with the questions of not just “Can we use AI in government?” but “Should we?”, our study of over 1,500 municipal employees in Austin set out to uncover practical strategies for integrating AI while upholding ethical governance.
We partnered with the City of Austin on a comprehensive research initiative to design and implement ethical AI training for city workers. By combining an extensive literature review, large-scale employee surveys, and hands-on workshops, we developed a replicable, ethics-first framework for government AI adoption that is already serving as a model for municipalities nationwide.
How We Designed Austin’s AI Ethics Program
We began with three primary research questions:
- What is the scope of AI tools and associated practices currently used by City of Austin staff?
- What are the most effective ways to structure, develop, and deliver AI training in city government?
- What curriculum and online resources are most valuable, and which content or instructional approaches work best for different user groups?
To address these, we reviewed 76 academic and organizational sources, surveyed 1,502 employees, and conducted five interactive workshops with 125 participants across key departments.
We administered both a general survey to more than 16,000 city workers and a targeted survey for data-centric roles. Workshops were delivered in both in-person and virtual formats between March and July 2025 to compare training effectiveness.
Current AI Usage Among City Workers
Survey data showed that 25% of city employees were already using AI tools such as ChatGPT, primarily for communication, content creation, and data analysis. Usage was notably higher among data-focused roles, with over half reporting regular engagement. While many recognized AI’s potential to boost productivity, most also raised concerns about errors, privacy, transparency, and algorithmic bias. More than 80% agreed that training is essential for safe and effective AI adoption.
What Employees Want from AI Training and What Really Works
Analysis of learning platform data and employee feedback revealed strong preferences for practical, scenario-based training focused on real workplace challenges and ethical AI use. Most respondents wanted instruction on how AI would impact their daily work and on broader safety considerations. Our “Ethics-First” workshop series combined critical questioning, case studies of real-world failures, and peer scenario exercises that addressed themes like accountability, transparency, and environmental impact. Feedback was highly positive, with participants valuing the interactive and context-specific approach. Notably, virtual sessions often promoted higher engagement than in-person formats. This shows that city staff aren’t just looking for AI skills; they want practical tools and clear frameworks to use AI responsibly in their day-to-day work.
Ethics-First Framework: Four Guiding Principles and Seven Key Issues for Responsible AI in City Government
©Image made with Figma
Building Sustainable Training Infrastructure
We drew on resources from the GovAI Coalition and designed adaptable training materials that can evolve with new policies and technologies. Key recommendations for effective city AI training programs include centering ethics in curricula, using interactive methods, supporting ongoing communities of practice, and updating content regularly in line with policy and technological change.
Implications for Municipal AI Adoption
Our study demonstrates that focused, ethics-driven AI training can build employee confidence, address key concerns, and enable responsible adoption in public service. This approach supports technological innovation while upholding democratic governance and public trust. The framework developed in Austin now provides a practical, adaptable model for cities nationwide.
Want to learn more?