The Detroit Police Department has proposed a new policy for using facial recognition technology, but it’s already opposed by a coalition of civil rights groups.
DPD has been using facial recognition without a formal oversight policy in place for more than a year. The department withdrew an initial proposed policy after the Detroit Board of Police Commissioners tabled a vote on it in June, and submitted a revised version in its place last week.
The new proposal puts limits on how police can use facial recognition software. For example, it can only be used on still images of criminal suspects linked to serious crimes, such as homicide and armed robbery. It prohibits using the software to scan faces in real time, though the technology the department has is capable of doing that.
The policy also pledges that DPD “will not violate the First, Fourth, and Fourteenth Amendments,” and prohibits the use of facial recognition based on “religious, political or social views” and races, ethnicities, and citizenship, among other things. It also lays out penalties for employees who violate the policy, stating that such breaches will be “investigated for criminality” and “the remedy for this misconduct is dismissal from DPD.”
But in a letter sent to police commissioners on Thursday, a coalition of civil rights groups warns the technology is still dangerous, and urges police commissioners to reject the policy.
“Facial recognition technology is flawed and dangerous, and its use in any form by the DPD can serve only to further erode trust between the DPD and Detroiters—especially Detroiters of color,” wrote the groups which include the ACLU of Michigan, the Michigan Immigrant Rights Center, and the Detroit Hispanic Development Corporation. “Detroit and the DPD should invest their time and resources in improving community relations and in making real changes in our neighborhoods to improve public safety rather than in finding new ways to use technology to police our neighborhoods from afar.”
The groups go on to say that their concerns cover two main areas: that facial recognition technology is “a threat to communities of color and immigrant communities,” and is “a unique threat to our privacy.” The letter cites studies that found facial recognition algorithms demonstrate higher rates of error identifying women and people of color, and are used by federal agencies such as U.S. Immigration and Customs Enforcement (ICE) “to target immigrant communities for enforcement actions.”
The DPD says facial recognition identifications are always verified by multiple trained analysts, are only used as investigative leads when verified, and can never be used as the basis for an arrest without corroborating evidence. The department also says it would never use the technology for immigration enforcement, though the policy contains no explicit statement about sharing information with other government agencies.
“We appreciate that the DPD’s current policy proposes not to use many of the surveillance capacities that it has already purchased,” the letter continues. “But the limitations the DPD now suggests it might be willing to accept are in significant tension with its decision to purchase technology with such sweeping capabilities in the first place.”
The letter ends by urging the Board of Police Commissioners to “deny the DPD the ability to deploy facial recognition technology in any form. Allowing facial recognition technology today sows the seeds of the surveillance state of tomorrow. We implore you to act now before these seeds can grow into something we can no longer uproot.”
The Board of Police Commissioners did not vote on the policy at its weekly meeting on Thursday. It’s unclear when they will do so.