See the issue brief: Detroiters' Cautious and Conditional Support for Artificial Intelligence in Local Government
DETROIT-Many Detroiters are still making up their minds about potential government use of artificial intelligence, with 41% of city residents reporting neutral views or uncertainty on whether AI's benefits to society outweigh its risks.
A new issue brief outlines findings from a representative survey of Detroiters fielded by the Detroit Metro Area Communities Study at the University of Michigan. More than 2,100 people responded to the survey between Aug. 6 and Oct. 1, 2025, and survey results are weighted to represent the demographics of the city's population as a whole.
The survey asked Detroiters about their general perceptions of AI as well as their support for specific municipal AI applications, including managing water and wastewater systems, assisting residents in finding government resources, monitoring neighborhood upkeep, identifying crime suspects and identifying missing children.
"We wanted to know whether residents think AI can improve efficiency, transparency and accountability in government services as well as whether the technology might increase harm, bias or unfairness in service delivery. We hope city officials will take Detroiters' views on AI into account as they explore applications of AI in government services," said Heonuk Ha, U-M postdoctoral research fellow at the Institute for Social Research and an AI fellow at the Institute for Data and AI in Society.
Ha co-authored the issue brief with Yucheng Fan, data manager at DMACS, and Mara Cecilia Ostfeld, a faculty co-lead at DMACS.
The survey found Detroiters' support for AI depends on the purpose for which the technology is being used. Support for AI was strongest for identifying missing children, where a majority of residents (57%) expressed approval. Support for AI is much lower for service-oriented applications such as monitoring neighborhood upkeep (38%), helping residents navigate government resources (34%) and managing water systems (30%).
Using AI to identify crime suspects received the highest level of opposition (39%). This pattern may partly reflect that 55% of residents agree that this use of AI could harm some residents even if it helped others.
"When we asked Detroit residents about the potential for AI to improve government efficiency, transparency, and accountability, the survey found efficiency was the most persuasive benefit of AI, but Detroiters still were not widely convinced," Fan said. "Between 20% and 30% of residents were neutral about efficiency gains in different scenarios, which suggests efficiency alone does not lead to broad public buy-in."
Detroiters' concerns about harm and bias related to AI are persistent. Across all use cases, large shares (40-55%) of residents agree that AI could harm some residents even if it helps others. Even for the AI-use case that was most supported-identifying missing children-concerns about bias remain substantial (41%).
Residents also are skeptical that AI will automatically produce more equal service delivery. Neutral responses dominate perceptions of fairness, with about 30% to 35% of respondents reporting that they neither agree nor disagree that AI will increase fairness and improve service delivery across use cases. Another 10% to 20% of Detroiters said they "don't know" whether AI would increase fairness in various types of government service delivery.
Support for AI drops sharply when Detroiters' personal data is involved. Nearly two-thirds of residents (64%) would not allow access to their personal data, and only 22% would be willing to permit access to their own personal data in government records to help improve the accuracy of a new tool. The other 14% don't know what position they would take.
"Detroiters are approaching the use of artificial intelligence in local government with caution rather than enthusiasm. Many residents remain unsure about the benefits of these technologies, and concerns about data sharing are particularly pronounced," Ostfeld said. "If the city hopes to expand the use of AI, it will be important to first build trust by clearly explaining how these systems will be used and how residents' information will be safeguarded."
