Can facial recognition be trusted by police? Detroit says no, with new rules to prevent wrongful arrests based on faulty technology.
Michigan, June 30, 2024:
The recent settlement between the city of Detroit and Robert Williams, a Black man who was wrongfully arrested due to faulty facial recognition technology (FRT), has sent shockwaves through the law enforcement community. This landmark decision places some of the most stringent restrictions on police use of FRT in the nation, prompting a crucial conversation about responsible technology use and safeguarding civil liberties.
Williams' story is a stark reminder of the dangers of relying solely on FRT. In 2020, an expired driver's license photo flagged by facial recognition technology misidentified him as a shoplifting suspect. This faulty match triggered a chain of events that led to his arrest in front of his family, a traumatic experience that "completely upended my life," as Williams wrote in a recent TIME op-ed.
This wasn't an isolated incident. Detroit PD has acknowledged at least two other wrongful arrests based on FRT, both involving people of color. Studies have consistently shown that facial recognition algorithms are more likely to misidentify Black, Asian, and Indigenous people. This inherent racial bias raises serious concerns about the potential for discriminatory policing practices.
New Rules for a More Just Approach
The Detroit settlement aims to address these issues head-on. The new regulations place significant limitations on how police can utilize FRT.
Gone are the days of relying solely on a blurry photo match for an arrest. Police can no longer use FRT results as the sole justification for detaining someone.
FRT can now only serve as a lead, requiring further investigation and the collection of independent, reliable evidence that definitively links a suspect to the crime. This additional hurdle helps prevent innocent people like Williams from being caught in a web of faulty technology.
Detroit PD officers will now undergo mandatory training that specifically addresses the documented racial bias of FRT. Understanding these limitations will equip officers to use the technology more responsibly and critically analyze its results.
A comprehensive audit will be conducted for all cases since 2017 where FRT played a role in obtaining arrest warrants. This audit could uncover potential past injustices and identify areas where investigations were flawed due to overreliance on FRT.
A Model for Responsible Tech in Policing?
The Detroit settlement holds significant promise. It has the potential to serve as a blueprint for other law enforcement agencies across the country grappling with the ethical dilemmas surrounding FRT use. By prioritizing independent evidence, mandating anti-bias training, and auditing past applications, Detroit is taking a critical step towards ensuring responsible technology use in policing.
However, challenges remain. The agreement only extends for four years, raising questions about its long-term impact. Furthermore, a complete ban on FRT is not in place. Ongoing public discourse, advocacy efforts by civil rights groups, and continued monitoring by external bodies will be crucial to ensure that these initial steps lead to lasting reform and responsible police use of technology.
The Detroit settlement signifies a turning point. It acknowledges the dangers of FRT when used without proper safeguards and prioritizes protecting individuals from potential abuse. While the journey towards a future where technology enhances public safety without compromising civil liberties is ongoing, Detroit's course correction offers a beacon of hope for responsible policing in the age of advanced technology.