‘Common Law’ Explores Bias in Algorithms

Algorithms using reams of data shape everything from insurance rates to parole decisions to whether you are screened more vigorously at the airport. As a set of instructions made by people, algorithms can bake in human bias, but also potentially counteract it, Professor Deborah Hellman says on the latest episode of “Common Law,” a podcast sponsored by the University of Virginia School of Law.

Hellman, an expert on equal protection and author of the book “When Is Discrimination Wrong?,” has written two papers on the subject. “Big Data and Compounding Injustice” is forthcoming in the Journal of Moral Philosophy and “Measuring Algorithmic Fairness,” was published in the Virginia Law Review and recently won the Association of American Law Schools Section on Jurisprudence Article Award. Hellman is director of UVA Law’s Center for Law & Philosophy.

In the podcast episode, Hellman explains that looking into algorithmic fairness has also influenced her thinking on when discrimination is wrong.

“Discrimination is wrong when it takes a situation which is already unjust and makes it worse,” Hellman says, pointing to the kind of algorithms that might lead an insurance company to charge a battered woman more on her life insurance premium.

Hellman explains different ways bias can compound injustice, for example by using data from performance reviews in an algorithm used to identify promising new recruits.

“Suppose that the people who are labeled ‘good employees’ are disproportionately men and the reason they’re disproportionately men is that the managers who’ve been writing performance reviews are biased against women,” she says. “So if that were to be the case, then the machine learning tool is going to kind of ‘learn’ that men are better employees than women. And what the algorithm is doing is it’s just carrying that bias forward.”

This way in which algorithms compound injustice has attracted significant attention.

“What has attracted less is the way that prior injustice produces real differences among people and the more we make new decisions based on facts determined by this unjust past, the more we compound injustice,” Hellman said outside of the episode. “The tantalizing power of algorithms may exacerbate this dynamic.”

On the other hand, Hellman emphasized, the use of algorithms offers an opportunity to intervene.  Perhaps algorithms could use “race” and other protected traits to help reach equitable results.

Focusing on “Law and Equity” for its third season, the show features Dean Risa Goluboff and Vice Dean Leslie Kendrick ’06 as hosts.

While the themes of the first two seasons were temporal — the first focused on “The Future of Law” and the second looked back at “When Law Changed the World” — this season looks across time at a variety of legal issues, asking what equity means and examining how it interacts with law.

“Common Law” is available on Apple Podcasts, Stitcher, YouTube, Spotify and other popular places you can listen to podcasts. The show is produced by Emily Richardson-Lorente.

You can follow the show on the website CommonLawPodcast.com or Twitter at @CommonLawUVA.