Since 2009, the University of Michigan Ford School of Public Policy’s Center for Local, State, and Urban Policy has conducted the Michigan Public Policy Survey, a biannual survey of local government leaders across the state. This year’s report, “Michigan local government leaders report significant increases in police surveillance technology, uncertainty about AI introduction,” was published in April by CLOSUP researchers Debra Horner and Margaret Walthall with support from Sydney Lount and Samantha Clark, Undergraduate Research Opportunity Program student researchers.
In an interview with The Michigan Daily, Horner said the MMPS is a census survey of local governments in Michigan, covering all 1,856 counties, cities, villages and townships across the state. In 2024, the survey expanded to include more perspectives on the selected policy issue and research focus.
“In 2024, we decided that our topic that we were going to research was policing and criminal justice issues,” Horner said. “We added on two additional surveys: one of sheriffs and chiefs of police, and one of county prosecutors. So, in spring last year, we did local governments, police chiefs and sheriffs and prosecutors to get the wide range of actors who work in policing and criminal justice issues to get their opinions on a whole bunch of different topics related to public safety.”
The report draws comparisons with data collected in the Fall 2015 MPPS report, highlighting an increase in the use of police surveillance technologies in local communities. Horner said there was significant growth in the use of existing community surveillance technologies such as dash cams for police officers, as well as an emergence of new technologies.
“We also asked about some new technologies, ones we didn’t even kind of know to ask about or think to ask about back in 2015 that we added to the 2024 survey,” Horner said. “Questions about automated license plate readers, so tracking people by their license plate, and facial recognition technology, which is a very controversial topic right now in terms of the accuracy of facial recognition.”
Research in the 2025 report found that 55% of local government officials reported being unsure whether automated tools are more or less accurate than human judgment, with 59% of police chiefs and 66% of county prosecutors believing the same. In an interview with The Michigan Daily, Lount, Public Policy rising junior, said she was surprised by the high uncertainty levels surrounding automated tools.
“For me, that was the most surprising and also the most concerning, in a way,” Lount said. “If we’re looking at this study and what types of technology that’s being used, and we’re seeing that police are saying we are using this type of artificial intelligence, but there’s a large percentage of local government officials, police chiefs and prosecutors that don’t actually know if AI is more accurate or not — that to me seems pretty problematic. It’s hard to shape policy and actually be using this type of technology if we are not 100% sure of the implications.”
Lount said there is also concern surrounding inherent human biases manifesting in artificial intelligence surveillance tools.
“AI is basically trained by humans, it’s built by humans, and humans inherently have bias,” Lount said. “So if humans are the ones that have bias, and they’re building these sorts of systems, these policing systems, then those (biases) can sort of work (their) way into the AI. And I think a lot of people are concerned about racial bias, or there could be socioeconomic bias, with where we’re choosing to police or matters like that.”
In an interview with The Daily, Kentaro Toyama, professor of community information, referenced his technology research where he coined this concern as the “Law of Amplification.”
“(This is) what I call the Law of Amplification of technology, which basically says that digital technologies’ main impact is to amplify underlying human forces,” Toyama said. “And what that means specifically is that whatever people are already doing, technology tends to emphasize. And in particular, you can’t use technologies to fix problems that are social in nature.”
Toyama also said there were potential conflicts AI surveillance technology could have with Fourth Amendment protections against unreasonable search and seizures.
“One of the things that might start happening is, as you have systems that can track license plates and know exactly where somebody is at any given point, an AI system could conceivably put together patterns that an actual human being would have never noticed before,” Toyama said. “(It could) then make inferences about what that person’s intention is, or whether they’re committing a crime or possibly committing something that looks like a crime based on the data, (even when someone) is completely innocent. I think that will raise a lot of legal and ethical questions, and that needs to happen.”
Horner said local government leaders have an opportunity to develop and implement safeguards surrounding the widespread use of AI and new surveillance technology.
“A really good place to look to put restraints or regulation toward the use of these police technologies is in the decision makers at the community level,” Horner said. “I think that’s a really good opportunity to have local boards, who are the people who oversee policing agencies, to develop formal policies around the use of these technologies and not just necessarily let it be adopted at random without a lot of forethought.”
Daily Staff Reporter Aanya Panyadahundi can be reached at aanyatp@umich.edu.
