Artificial Intelligence (“AI”) is increasingly taking over our world by the day. There is much on going research and speculation about how AI through automation and augmentation will change our world. Irrespective of whether one predicts a utopian or dystopian future, the nature and frequency of crimes committed would have drastically changed by the end of the next decade. There is, however, a fault line in the present system which cannot be overlooked.
In 2016 the Supreme Court of Wisconsin convicted a man named Eric L. Loomin for a number of charges, and sentenced him to six years imprisonment based on a recidivism assessment done by a proprietary closed source software named COMPAS. When appealed to the United States Supreme Court, on the ground that the defendant’s Constitutional right to due process of law for trial was violated, the court declined to hear the case.
A person’s Constitutional right to due process of law safeguards him/her from arbitrary denial of life, liberty or property by the government outside the sanction of law. The Constitutional right to due process of law, in this case, was violated because the legal reasoning required to determine recidivism, which is an important discretion of a judge was replaced by COMPAS. A Judge needs to weigh various factors such as, nature of the offence, criminal history, scope of clemency, etc to determine the same. This responsibility is not mathematical in nature and requires strong legal reasoning and legal ethics; therefore when the same case is presented to two different judges they may reach different conclusions based on strong set of reasons whilst upholding the constitutional principles of the jurisdiction.
On closer look, Artificially Intelligent softwares like COMPAS are created and fed by human beings and big data and mostly run on algorithms which is ultimately derived from human behavioural patters, biases and even strong regional stereotypes. AI is therefore not free from bias. Keeping in mind that a Judge, being human, also has some bias, still does not put AI on the same ground as that of a judge, because AI ultimately is a magnified product of human biases. Moreover, in this case if COMPAS was an open source software it would have been worse for the person in question, because then the person’s fate would be based on the statistics of people contributing to the software big data.
If the day ever comes where AI replaces Human intelligence, judiciary would not be independent anymore, it would in fact become dangerously popular. In such a scenario, reason would be superseded by persuasiveness of media to the mass, and justice would be reduced to a popularity contest.
In conclusion, Artificial Intelligence revolution, just like the last Industrial Revolution is set to change our lives, demographics and environment forever and maybe even solve some of the most crucial issues of our times. However, no level of Artificial Intelligence can replace Human Intelligence; it can merely enable us to make better decisions.
A word of caution: The fact that AI cannot replace human intelligence does not mean it will not drastically modify the way we practice our respective professions, for better or for worse.