March 3, 2023 | OPINION | By Pierce Sullivan, Sports Editor
In 2015, a Quebec judge asked a woman in his courtroom to remove her hijab. He was convinced that he needed to be able to see her face in order to be able to read her. The judge felt that the hijab inhibited his ability to make an accurate assessment of the case at hand. If he could not see her face and read her expressions, then he did not feel he could be an impartial, accurate judge.
We, as humans, have a profound and unwavering faith in our ability to read people; in our ability to discern things about the nature of a person from their mannerisms and facial expressions.
The Quebec judge has since apologized for his actions, so, for the sake of this argument, I give him the benefit of the doubt. Maybe he wasn’t just a racist bigot on a power trip, but instead truly did believe that he needed to see the defendant’s face to make an accurate decision. Because he, as a judge, must possess an innate ability to see past the superficial and unveil the truth.
It is a judge’s job to set bail; thus, it is a judge’s job to determine who is at most risk to commit another crime in the future. If that is their job, they better be damn good at it.
In his book “Talking to Strangers”, Malcolm Gladwell discusses a phenomenon surrounding this situation. Gladwell brings up a 2017 study, which compared the efficacy of an AI judge and real judges.
The researchers gathered records of 554,689 defendants from New York City, of which 400,000 had been previously released by judges.
The AI judge was then fed the same information that the judges had been—criminal records, demographic backgrounds, and details of charges—and was programed to provide a list of 400,000 people to “release.” Everything was the same. Except for the judge’s remarkable ability to read people and see the truth, of course.
The two lists of 400,000 released defendants were then compared. The AI generated list was 25 percent less likely to commit a crime while on bail. In the world where the human judge makes the call, that’s 100,000 people in prison who should not and do not need to be. That’s our world.
Our biases do not make it easier to make judgements. They drastically skew our perspective on the truth. When I say, “I welcome our A.I. overlords and you should too,” this is not hyperbole. It has been proven that artificial intelligence does a better job at judging human character than a person can.
The implications of this are profound. On the most basic level, this could be a first step in reforming the United States’ deeply flawed criminal justice system. A.I. provides a way to entirely avoid racial biases in the courtroom.
Some may voice arguments against the use of A.I. in a courtroom, citing fears of letting important judgements free from the binds of human control. But I question whether those who fear this are being genuine, or if their fear lies more in the implications for humanity.
A.I. technology has grown by leaps and bounds since this 2017 study. As it becomes more sophisticated, it will only grow in accuracy. A.I. systems have proven that they have the ability to watch over us better than we humans are able to.
It is time to welcome them with open arms.