Facebook remembers what products you were looking for last week. Google knows what to advertise in search results. Netflix makes suggestions based on your viewing preferences. We experience big data and the results of data analytics nearly every day. Now it’s coming to an HR Department near you.
No doubt big data is full of promise for HR: recruiting faster and smarter, knowing when your trade secrets are sneaking out the door, and identifying management deficiencies are among the many issues that data analytics can help employers address. These are exciting uses – things that employers want, and, in some cases, need to improve their businesses. I don’t want to discourage employers from using big data to reaching their goals. I just have to ask: Have you thought about discrimination?
When I raise the possibility that using big data to make human resources decisions can lead to a discriminatory disparate impact on members of a protected class, the reaction I get is usually a blank stare. Sometimes the response goes like this: “Using big data can’t possibly be discrimination. It doesn’t involve people. It’s just a computer program crunching data, using an algorithm. The program doesn’t know anything about anyone’s protected class status!” That’s true, but it doesn’t get the use of big data off the hook. Disparate impact discrimination has nothing to do with intent. Plaintiffs (or classes of them) do not have to prove intent to prove discrimination. They only need to show that an employer engaged in an employment practice that had an unjustified adverse impact on members of a protected class. Just look at three of the top ten discrimination settlements of 2013. In these cases, the discriminatory impact of seemingly neutral policies resulted in seven-figure payments to plaintiffs, without a showing of intent by the employers.
So how can an employer know if its use of big data could have a discriminatory disparate impact? One way is called validation. Validation has been used for decades to evaluate employers’ use of testing (or “other selection procedures”) when making employment decisions. Companies (the EEOC and OFCCP) “test the test,” using statistical benchmarks to see if protected groups are disadvantaged. The same approach can be used to support the legality of personnel decisions based on big data and data analytics and, if necessary, to defend claims of disparate impact.
Worry-wart management-side employment lawyers are not the only ones who think about discrimination and big data. The White House and privacy professionals are concerned too. Can plaintiffs’ employment lawyers be far behind?
Posted by: Kate Bischoff
Posted by: Kate Bischoff