Survey researchers characterize a ‘Wild West’ period, express optimism for transparency, regulation
As the utilization of surveillance technology by law enforcement has surged, several officials in Michigan remain cautious about the integration of newer AI-driven policing instruments, according to a survey from the University of Michigan.
The Michigan Public Policy Survey, carried out last spring by U-M’s Center for Local, State, and Urban Policy, revealed that a majority of officials are uncertain regarding the precision of public safety risk evaluations conducted by automated systems like AI. Nevertheless, sheriffs and police chiefs generally have a more favorable view of AI evaluations compared to local government representatives and county prosecutors.
A third of law enforcement leaders in Michigan state that their agencies have implemented or are likely to adopt predictive tools in their policing strategies.
The implementation of more familiar policing technologies has markedly increased over the past decade. From 2015 to 2024, the reports of local law enforcement agencies using body cameras jumped from 23% to 77%, dashboard cameras from 55% to 74%, public surveillance cameras from 28% to 50%, and drones from 3% to 31%.
However, the recent integration of advanced technologies is less pervasive. In 2024, merely 26% of sheriffs and police chiefs indicated that their agencies utilized automated license plate readers, 10% adopted facial recognition systems, and just 3% employed AI or other predictive instruments.
This reserved attitude may illustrate apprehensions regarding the reliability of automated tools. More than half (55%) of local government officials expressed uncertainty about whether these tools are more or less precise than evaluations made by humans. Sheriffs and police chiefs (59%) and elected county prosecutors (66%) demonstrated even deeper skepticism.
“It’s understandable there’s considerable ambiguity around the precision and repercussions of tools like predictive policing—most communities lack firsthand experience with them thus far,” said Margaret Wathall, research area specialist on the survey. “These technologies are still emerging and lack established regulations or verified records of success. As these tools evolve and more communities become familiar with them, we hopefully will see enhanced clarity regarding their responsible use.”
Opinions varied on whether new technologies or equipment are beneficial. While both factions recognized the merit in such investments, 85% of law enforcement leaders “strongly agreed,” in contrast to only 62% of local government officials.
The survey additionally indicated differing levels of assurance in AI’s application within prosecutorial functions. Half of Michigan’s county prosecutors expressed confidence in AI applications for identifying high-risk neighborhoods, while 45% endorsed its use for processing and scrutinizing forensic evidence.
Trust was lower regarding AI utilization for monitoring social media for potential criminal activities (38%) and drafting legal documents (25%). Prosecutors exhibited the least confidence (20%) in AI’s ability to conduct evaluations for sentencing, parole, probation, and release.
“AI is still navigating a somewhat Wild West phase,” remarked Debra Horner, senior project manager on the survey. “However, over a third of sheriffs and police chiefs indicate they are either currently employing or soon to implement predictive tools for policing, and nearly half of prosecutors report at least some level of trust in AI or other automated applications for certain functions.
“This indicates it may be an opportune moment for the state of Michigan and individual local governments to contemplate establishing formal regulations surrounding the transparent and responsible application of AI in law enforcement.”
CLOSUP, which is a part of U-M’s Ford School of Public Policy, received survey responses from county, city, township, and village officials from 1,307 jurisdictions across the state during April-June 2024. Additionally, responses were collected from 54 county sheriffs, 234 police chiefs or public safety directors, and 55 elected county prosecutors.
The spring 2024 iteration of the survey program was supported by a grant from The Joyce Foundation.
Authored by Rebecca Cohen, Ford School of Public Policy