If you read the news (which we understand can be a little overwhelming at this point), then you already know that the use of facial recognition software is a highly controversial subject. Here in Michigan, the Detroit Police Department has been using Facial Recognition Software for years, which has led to some highly publicized instances of mis-identification. According to the National Institute of Standards and Technology, when they reviewed the industry’s leading facial recognition software algorithms, they found them to be more than 99% accurate when matching high quality headshots to the database of other full frontal poses. But how often does a police officer have a high quality headshot of a suspect on hand for comparison?
A recent case in Detroit illustrates this point very well …
Robert Williams, a resident of Metro Detroit, was arrested in 2018 by the Detroit Police after he was incorrectly identified as a suspect in a shoplifting incident involving stolen watches. Williams says police identified him as a suspect from video footage of the theft, which was then matched to his face using Facial Recognition Software. According to the DPD, that was back when they were using the Michigan State Police’s facial recognition software. Once they got their own system, the error rate decreased dramatically because they put new policies in place governing the use of the software in investigations. But is that enough?
There is a history of facial recognition software misidentifying people…
The Robert Williams case from Detroit is only the tip of the iceberg. In fact, there have been numerous cases where people have been misidentified based on their skin color, as the software is widely known to struggle when identifying people of color. (Who could forget the controversy in 2015 when Google issued a formal apology after it’s then-new “photos” software labelled several black people as “gorillas.” Since then there have been numerous instances, many of them related to the criminal justice system. As experienced criminal defense attorneys defending people in Grand Rapids, Lansing, Detroit and Howell, we’ve certainly seen our fair share of people wrongfully accused of crimes!
The current controversy includes Facebook and the feds
In January of this year, Facebook agreed to pay $550 million to settle a class action lawsuit surrounding its use of facial recognition software for data mining purposes, which violates Illinois’ biometric privacy laws. As for the federal government, a police reform bill was recently introduced by the House of Representatives that would limit the use of facial recognition software by law enforcement agencies around the U.S. For example, police departments making use of federal grants to buy bodycams would have to adopt policies on their use of facial recognition on any footage collected from the body cams. This would include getting a judge’s approval, and only using it when there is an “imminent threat or a serious crimes.”‘
Being falsely identified can lead to wrongful convictions!
Join us next time as we wrap up this discussion about facial recognition software, and look at the controversies surrounding its use, both here in Michigan and across the country. Until then, we’d like to take this opportunity to remind our readers that the best way to avoid being wrongfully convicted of a crime is to hire the best criminal defense attorney available to you! And by ‘best’ we mean the most experienced, trusted, educated, aggressive, and with access to some of the best expert witnesses in the field that can help defend you against false allegations and wrongful convictions. We’ve been doing this for over a quarter century, and we’re very good at it. So whether you live in Kalamazoo, Mt Pleasant, Midland or Lansing, call 866 766 5245 today and make sure you’re properly protected.