Ver la versión en español aquí

COVID-19 has thrown a wrench into every employer’s near-term plans. While many companies have temporarily shut down, and many others have put hiring on hold, some employers will have to kickstart their hiring efforts as the economy begins to reopen. Data shows that the stay-at-home economy created a surge in social media usership, so it will be no surprise when home bound recruiters try to use Facebook and other platforms to reignite their workforce. HR professionals who currently use, or plan to use social media advertising for online recruiting need to understand its potential legal pitfalls.

The charity Lifewater International posted an online Facebook advertisement for an engineering position whose role would support aid efforts in Africa and Southeast Asia. All it wanted was for the ad to reach qualified applicants who were eager to serve the world’s poor. Instead, the ad was viewed by 70% men mostly between the ages of 18 and 34. This happened even though Lifewater did not specifically target any demographics. Similarly, when the family-owned plumbing company, Barker & Sons, placed employment advertising on Facebook without any specific targeting, the advertisement’s viewership was 100% men.

Businesses, small and large, use Facebook’s advertising reach to find the best candidates for their open positions. But what these employers are discovering is that the artificial intelligence (AI) behind Facebook advertisements inadvertently discriminate by gender and age.

Many people know that Facebook has been under fire in the recent past for its targeted employment advertising. That was supposed to change in March 2019 when the company reached a settlement of several lawsuits alleging discriminatory job advertising. As part of the settlement, Facebook vowed to create a new advertisement algorithm that would ensure that its employment ads did not discriminate. However, new research by Northeastern University suggests that Facebook has not delivered on last year’s promise.

The new reports show that despite Facebook’s removal of demographic features from its ad delivery system, the new job-advertising algorithm still appears to discriminate based on gender and age. Further, the study demonstrates that diverse online audiences are not always found on the other end of a Facebook ad, even if an employer is purely seeking the best candidate.

We’ve discussed the failure of recruiting algorithms on LinkedIn and other services in the past, and it seems Facebook and its competitors are still a few years away from delivering non-discriminatory online recruiting. So even in 2020, the humans in your HR department (even when shuttered at home) are still the best bet at finding the best candidates. The stay-at-home economy has increased employers’ reliance on AI for job advertising. Here are a few tips to mitigate your risks:

  • Review your job advertisements’ viewership data and remove ads that disfavor protected classes. If the data shows that your ads are only being seen by a certain group of people re-evaluate whether this form of advertising is worth the risk of litigation and consider a new recruiting strategy.
  • Audit your advertisements’ content and remove any language that may skew viewership away from a protected class. For example, the terms ‘salesman’ or ‘waitress’ may result in gender bias in an online advertisement algorithm.
  • Do your research! Be sure to pick online advertisers who can reach the best candidates for the job irrespective of demographics.
  • Consult an employment attorney if you have concerns that your online recruiting strategy could be conflicting with discrimination laws.

*Special thanks to Thomas Raine, who assisted in the drafting of this post. Thomas is a third year Juris Doctor Candidate at the University of Miami School of Law.