Death Penalty

While the death penalty is inherently irrevocable and unjust to those sentenced to death, the variables that lend themselves to determining death sentences all become complex indicators in deciding who will live and who will die.

Even though one person  executed would be one too many, the actual number of people sentenced to death and the number of those that are actually executed make up a minority of those who commit capital crimes. This makes the death penalty a sort of lottery, with race, legal representation, and prosecutorial misconduct being key indicators.

Two professors in criminology and computer science developed a software program that can predict which prisoners will be executed. To do this, they developed a neural network, capable of scanning data for patterns. They gave the program 1,000 cases in the U.S., tracked from 1973-2000, and focused on characteristics of the inmate--race, age, marital status, location, and the type of offense. No actual details about the case or legal representation were included.

After the program was given these statistics, the researchers then gave the program an additional 300 cases, in which they did not give the outcome of the case. As the program's neural networks had been trained to replicate what would bring about an actual execution, they were able to predict which prisoners had been killed with a 90% success rate.

Since the death penalty's reinstatement, states have continued to erratically apply the death sentence. Although it is supposed to be reserved for the "worst of the worst", the most unlucky are the first to die. In 1994, U.S. Supreme Court Justice Harry Blackmun declared that

 "Twenty years have passed since this Court declared that the death penalty must be imposed fairly, and with reasonable consistency, or not at all, and, despite the effort of the states and courts to devise legal formulas and procedural rules to meet this daunting challenge, the death penalty remains fraught with arbitrariness, discrimination, caprice, and mistake."

Blackmun's words still seem to be evident today in the way that the death penalty is applied in America. The computer program was still able to accurately predict a death sentence, despite knowing no facts about the case and outcomes, which only helps to illuminate why the application of the death penalty in the United States is so unfair.

The researchers are still experimenting with the system by trying different combinations of variables to account for which are the most influential. You can read the full article here.

Emily

DPAC