What’s Really Making Us Uncomfortable—The Use Of AI In Evaluating The Likelihood Of Recidivism, Or The Policy Of Sentencing Based On The Likelihood Of Recidivism? – Criminal Law

United States:

What is really making us uncomfortable – using AI in assessing likelihood of recidivism or the policy of conviction based on likelihood of recidivism?

June 09, 2021

Duane Morris LLP

To print this article, all you need to do is register or log in to Mondaq.com.

In his recently published New York Review article “Sentenced by Algorithm”, a review of the book by former SDNY judge Katherine Forrest, “When Machines Can be Judge, Jury and Executioner: Justice in the Age of Artificial Intelligence,” the current one rated SDNY judge Jed Rakoff highlighted the many shortcomings in existing AI products designed to predict recidivism rates for past offenders. These products are designed to guide judges in deciding whether to extend a defendant’s sentence based on a theory of “incompetence” – essentially to protect the general public from the possibility that the defendant will continue his criminal pattern in the future. As Rakoff explains in a nutshell, the products currently available have unacceptably high error rates, which usually tend to overly predict future crime. In addition, their “black box” design raises concerns about the assumptions underlying the algorithm and the defendant’s ability to effectively challenge the algorithm’s output.

The book review is informative and certainly an interesting introduction to the use of AI products in law enforcement. But perhaps the most salient point made by Judge Rakoff is made at the very end of his post: “In a broader sense, the fundamental question remains: even if these algorithms could be made much more accurate and less biased than they are now, they should be used in the criminal justice system Deciding who should be locked up and for how long In my opinion, it is fundamentally unfair to increase the prison sentence of a defendant on suspected future crimes. “

The idea of ​​imprisoning someone for a crime they did not commit and may never commit is inherently unsettling. And when the decision is separated from human judgment and empathy, it somehow feels even less fair, perhaps because of our innate distrust of what we cannot understand.

Relapse prediction AI products may not be well developed yet, but if the current evolution of AI in general is an indicator, these products could very soon become more sophisticated and accurate – and by and large almost certainly more accurate than any one human judgment. Then we have to ask the real question – the difficult question – that Judge Rakoff is raising.

Disclaimer: This warning has been prepared and posted for informational purposes only and is not offered as legal advice, nor should it be construed as such. For more information, see the firm’s full disclaimer.

POPULAR ARTICLES ON: United States Criminal Law

Fight against corruption 2021

Foley Hoag LLP

This is the third installment in our First 100 Days series, which examines key trends in business law and investigations in the early days of the Biden administration.