Can AI remove bias during recruitments?

Several tech startups are toying with the idea of AI-powered recruitment solutions, but can it really remove the biases that humans struggle with?
9 August 2018

The recruiting process is full of biases. Don’t you agree? Source: Flickr / Amtec Photos

When you’re looking for candidates, you often tend to make a choice based on your “gut” feel – which you believe is just your experience talking. But is it really?

Experienced recruiters train themselves to avoid the subconscious biases that creep in during the initial phases of recruitment, especially when they’ve got hundreds of applications for a single job opening.

“A desire to recognize and overcome bias is an important step forward. But even with the best intentions, hiring decisions can still be influenced by personal factors we may not even be aware of,” said Dell’s Chief Diversity & Inclusion Officer Brian Reaves.

Unfortunately, the bias has quite significant results. On candidates as well as on the company.

To remedy the problem, businesses are looking at using artificial intelligence (AI) to screen the scores of resumes that applicants send in.

What can AI offer to recruiters?

Well, AI makes it easy for businesses to screen resumes. As a result, recruiters become more efficient. However, there’s way more to it than meets the eye.

Ideal, a company that offers an AI-based recruitment solution claims that AI ignores demographic-related information such as the candidate’s name, race, sex, and age during screening and shortlisting.

As a result, all recruitment processes are automatically Federal Contract Compliance Programs (OFCCP) into the Equal Employment Opportunity Commission (EEOC) compliant.

According to Ideal, resumes with white-sounding names receive requests for interviews 50 percent more often than identical resumes with African American-sounding names and resumes with English-sounding names receive requests for interviews 40 percent more often than identical resumes with Chinese, Indian, or Pakistani names.

It is biases like these that an AI powered recruitment solution aims to help recruiters avoid.

In fact, in the future, more AI solutions are expected to help human resources (HR) teams to make employee appraisals more transparent and free from bias.

“While recruitment is a current focus to alleviate bias, as companies continue to implement technologies, they can begin to track a candidate’s entire journey through the company. This will provide insight into all stages of the career life cycle, and identify additional areas where bias may be impacting a person’s day-to-day responsibilities and potential for advancement,” hopes Dell’s Reaves.

But will we teach AI our biases?

The fact is, in the present state, AI doesn’t learn on it’s own. It’s an algorithm that people train.

When it comes to human resources, what’s the data that is available? Past job openings, applications for those openings, and successful hires.

Building a model based on that data, given the bias that crept into each of those hires, can be dangerous. In essence, we could be teaching AI to repeat our mistakes (in perpetuity).

“Personal prejudices can quickly become reflected in AI. In recruiting — a space in which sensitive and life-changing decisions are made all the time and in which we accordingly have established strong civil rights protections — these forms of vicious algorithmic bias are especially important to detect and act against,” Dipayan Ghosh, a Harvard fellow and former Facebook privacy and public policy official who is now with the New America think tank told the CNBC recently.

As things stand right now, it seems like AI isn’t ready to take over recruitments. However, tomorrow might bring something entirely new and exciting.