As artificial intelligence and machine learning have entered into our daily lives, nothing has been more hotly debated than facial recognition. Governments and law officials claim we need facial recognition in surveillance cameras as a method to fight crime, while the public has pushed back, stating the technology infringes on freedom rights and ethics.
Regardless of the opposition, law enforcement in many countries has implemented their use, including the USA and China.
Now, following a serious blunder by US law enforcement, a lawsuit has been filed in the city of Detroit for the first case of a wrongful arrest as a result of facial recognition technology.
Robert Williams claims he was not even through his front door when police made the arrest, suspecting him of shoplifting $4,000 worth of watches in an incident over a year prior.
“I came home from work and was arrested in my driveway in front of my wife and daughters, who watched in tears, because a computer made an error,” Williams said in a statement.
“This never should have happened, and I want to make sure that this painful experience never happens to anyone else.”
The arrest happened after some video footage captured at the scene of the crime was run through facial recognition software. Despite the footage being relatively ambiguous – with poor lighting and the shoplifter never actually looking directly at the camera – the facial recognition technology returned Williams as a match. This match, supposedly, directly led to the arrest of Williams.
However, since the arrest, the police involved have backtracked and admitted that “the computer got it wrong”, with Williams only being arrested because of a false likeness to the perpetrator.
“As a result, Mr. Williams was arrested without explanation on his front lawn in plain daylight in front of his wife and children, humiliated, and jailed in a dirty, overcrowded cell for approximately 30 hours where he had to sleep on bare concrete—all for no reason other than being someone a computer thought looked like a shoplifter” states the lawsuit, which was filed on Tuesday, April 13.
This marks the first time facial recognition has been blamed for the wrongful arrest of a person. Local law enforcement has since disputed the claim, stating that it was in fact not the facial recognition, but “just bad detective work”.
"Facial recognition was used, but that's not why the arrest was bad." Said Detroit Police Chief James Craig in a statement, reports The Detroit News.
Facial recognition is currently being used in everything, from law enforcement to unlocking your phone. It involves deep learning algorithms to correctly identify faces in real-time by analyzing distances between facial features, shapes, and specified nodal points to create a unique numerical signature for each person. However, facial recognition – like all AI – is currently still in its’ infancy, and is prone to being wrong.
This appears to be heightened when the technology tries to identify black faces apart. Current iterations of facial recognition show a significant drop-off in accuracy when trying to identify people with darker skin tones, showing a clear racial inequity that means it should probably not be used in law enforcement just yet.
Williams’ attorneys are claiming this may have happened here, and that Detroit should consider following others cities’ footsteps in banning the technology.