Is AI making recruitment less fair?

Due to the pandemic, the associated social distancing, and lockdown rules, companies had to fundamentally transform how they recruited new talent.

Rather than calling candidates into the office for face-to-face interviews, employers had to rely on tech tools like Microsoft Teams or Zoom to do virtual interviews. It is obvious that without tech companies would have struggled to stay afloat during COVID-19.

But is technology always a net positive in recruitment?

In a recent The Chad and Chase Podcast episode, Chad Sowash and Joel Cheeseman sat down with US Equal Employment Opportunity Commission (EEOC)’s Keith Sonderling to discuss the dark side of using technology, and particularly artificial intelligence (AI), in recruitment.

Sonderling starts by talking about his route from being an employment lawyer to becoming commissioner role at the EEOC, as well as how the commission is actually the federal body most involved in “modern-day issues that affect employees on a daily basis”.

It is not the Department of Labor or the National Labor Relations Board who deal with issues around “rights in the workplace”.

AI and recruitment

Sonderling tells Sowash and Cheeseman about how commissioners at the EEOC really champion specific workplace issues – and for him it is AI.

This is because AI is being used in the workplace, and particularly recruitment, right now, “so the conversation needs to happen now”.

Sonderling notes that many companies “want to genuinely diversify their workforce and take out some of the bias in recruiting. I like tech, I really think tech can help eliminate a lot of bias”.

But for it to do this, companies and the government need to work together to create a “standard” that means AI does actually help to improve diversity at work, and not the opposite.

To achieve this, it is important to ensure that the data going into the algorithm hasn’t been compromised by human biases.

Sowash says: “I think most people misunderstand that the decision AI is making doesn’t stem from AI itself, rather that it stems from human decisions [and] humans are biased.

So when “the human being program[s] bias into systems, and in this case, AI, your bias could impact thousands” of people’s lives. Sonderling responded: “You just summarized it perfectly”.

Sonderling adds that it is not that AI is discriminatory or misogynistic, it is all about the data being fed in; “the bias inputs give you bias outputs”.

Oversight of AI in recruitment

Therefore, employers need to be looking at what is being fed into the algorithm – they are liable, according to Sonderling with his lawyer hat on. “There is no question…that the employer using these tools to make decisions will be liable for the outcomes” that are discriminatory.

Cheeseman points out that it is “messed up” that employers often blame vendors, so he calls on employers  to “look in the mirror” and be extra careful when choosing their AI-based tech tools.

However, Sonderling notes that candidates are unsure about how to complain about discrimination when it is the result of AI. This explains why the EEOC, for instance, is not seeing lots of legal cases around AI-based discrimination at work.

But as AI becomes more important in hiring decisions, “that’s just something that all of us can’t ignore anymore,” according to Sonderling.

It is the job of the Department of Labor and EEOC to properly educate “employees about their rights and…employers about their obligations under the law”; for Sonderling, this is even more important and successful than enforcement.

“Let’s prevent the discrimination happening in the first place”, especially since there is such a desire to better understand technology and make it work to the benefit of everyone.

Whereas “enforcement should be on bad actors” who are “going to design AI to intentionally discriminate”, according to Sonderling. But ultimately Sonderling is clear that he doesn’t want over-regulation of AI at work; “there are so many benefits of using technology in the workplace and I want to see it flourish”.

He continues: “If you want to do [AI recruitment] the right way, you have unlimited tools available to you” from the EEOC and elsewhere. “But if you want to do it the wrong way, I also have other tools available to you, which you’re not going to like”.

Check out the full podcast to find out more gems of insight from Sonderling, Cheeseman and Sowash about AI in not only recruitment, but at work in general.

The Evergreen Podcast Network is a trusted partner of UNLEASH Podcasts.