The Michigan State Police runs the Statewide Network of Agency Photos (SNAP), a database of 4 million mug shots and 41 million driver’s license and ID photos from the Michigan Department of State. The FBI can request searches of at least 35.6 million of Michigan’s driver’s license and ID photos (GAO). African Americans are likely overrepresented in SNAP; they’re arrested at a rate 136% higher than their state population share.
Officers can run searches from a desktop computer or a mobile device. Trained examiners run desktop searches and potential matches are peer reviewed; results from mobile searches are not peer reviewed (011467–011468). MSP’s face recognition use policy, which it has made public, requires that an officer have probable cause before running a mobile search. Desktop searches are only required to be for a “law enforcement reason” (016824). Searches can be run to identify criminal suspects, witnesses, bystanders, victims, unknown decedents, and the incapacitated (010928, 011345, 016824).
MSP does not use face recognition with real-time video. In accord with Michigan law (MCL 28.243.2), the MSP deletes from its database mug shots and fingerprints of people who are arrested but never charged or are ultimately not convicted. Unlike any other agency in our survey, the MSP provided documentation that their audit regime was functional.
The MSP system uses Cognitec and NEC face recognition algorithms.
Sources and Notes: Michigan State Police, GAO, U.S. Census MCL 28.243.2 (Last updated: September 2016). You can review our scorecard criteria in the Methodology section. Numerical citations, e.g. (123456), refer to official records available by clicking "View Documents" below.