Algorithms at Work

Use of mathematical tools examined by ILR faculty
Thursday, March 1, 2018

Some might think of algorithms as most useful in marketing, but their use throughout the employment process, from hiring to promotions to training to exit interviews, is booming.

This was the topic of Tuesday’s webcast with the ILR faculty who will teach “Algorithms at Work” May 4-5 at the Cornell Tech campus in New York City.

While these mathematical tools can avert some human bias in decisions, at the same time, bias and unethical practices can inadvertently be built into them, said Ifeoma Ajunwa, assistant professor of organizational behavior at Cornell ILR and an associate member of the Cornell Law School faculty.

An algorithm — any set of instructions done repeatedly — can use machine learning or artificial intelligence to make predictions.

“But it doesn’t mean those predictions are useful,” noted Martin T. Wells, the Charles A. Alexander Professor of Statistical Sciences at ILR.

While some software vendors are carefully testing their products, the responsibility for ensuring algorithms aren’t discriminatory lies with human resources professionals and businesses, Ajunwa said.

Auditing of the results might be needed on a continual basis, similar to the financial model whereby businesses do an audit every year to make sure practices are proper, said M. Elizabeth Karns, senior lecturer in social statistics at ILR.

Continual audits of hiring algorithms is a business practice proposed by Ajunwa in a forthcoming law review article and in a talk that she gave at the Tech Law Colloquium at Cornell Law School in November.

The upcoming course will give participants the tools to understand how algorithms and machine learning work and how audits can be done to evaluate compliance, Ajunwa said.

Those questions are pressing. Nearly all Fortune 500 companies use online applications, and companies such as Goldman Sachs are moving toward automated hiring via video interviews and questionnaires,  Ajunwa noted. Algorithms are transcending all workplaces — blue and white collar alike.

Some lawsuits have resulted. Ajunwa pointed to Kyle Behm v. Kroger Co., currently being litigated. It alleges that a personality test the company used to filter out job applicants violates the Americans with Disabilities Act.