We are reminded of this obvious fact again with the news Friday that an innocent man, despite not looking like the perpetrator at all, was arrested last year after being falsely identified by faulty facial-recognition tech.
This is the second known case of facial recognition software directly leading to the arrest of an innocent man. It’s something privacy advocates fear will be a growing trend unless drastic action is taken to stop this technology in its tracks.
Michael Oliver, then 25, was charged with a felony for supposedly grabbing a phone from a car passenger and throwing it, reports the Detroit Free Press. Except, of course, it wasn’t Oliver. With facial-recognition technology’s demonstrated bias when it comes to identifying the faces of Black people (and BIPOC in general, and women, and old people, and young people…) it should come as no surprise that both Oliver and the actual phone-grabber are both Black.
The similarities stopped there, however. As the Detroit Free Press points out, Oliver’s arms are covered in very visible tattoos. In the video, the man who grabbed the phone is obviously tattoo-less. Still, Detroit Police arrested him anyway.
Dan Korobkin, legal director of the ACLU of Michigan, released a statement condemning both the arrest and Detroit police’s use of facial-recognition technology.
“Lawmakers must take urgent action to stop law enforcement use of this technology until it can be determined what policy, if any, can effectively prevent this technology’s harms,” the statement, emailed to Mashable, reads. “At the same time, police and prosecutors nationwide should review all cases involving the use of this technology and should notify all individuals charged as a result of it. This technology is dangerous when wrong and dangerous when right.”
Oliver’s case, while upsetting enough on its own, isn’t even unique — a fact that makes it all the more terrifying. Late last month we learned that Detroit police arrested a (yes) Black man by the name of Robert Julian-Borchak Williams after facial-recognition software sloppily matched his driver’s license photo to blurry surveillance video.
“I guess the computer got it wrong,” a cop reportedly told Williams after he had already spent 30 hours in jail.
Detroit authorities, for their part, insist that this was a one-time — er, two-time — mistake, and that it could never happen again.
“As a result of these two cases, we have a more stringent protocol in facial recognition cases,” Wayne County (the county in which Detroit falls) prosecutor Kym Worthy told the Detroit Free Press. “The cases will be reviewed during the warrant charging phase, prior to the preliminary examination, and again when the case is bound over to the Circuit Court in any case where facial recognition has been used as an investigative tool.”
There, don’t you feel better? The next time a person is arrested for a crime they didn’t commit based on some garbage facial-recognition algorithm riddled with errors, authorities will at least have gone through a more stringent protocol first.