Artificial intelligence is taking over some parts of the recruiting process that used to be reserved for humans.
Take for example, HireVue’s software that analyzes facial expressions and word choice in video interviews. Koru’s does the same with written tests, while Fama and TalentBin scour social media to profile candidates. RoundPegg uses automation to determine a candidate’s potential cultural fit with a company. There is even an AI chatbot named Mya that can help candidates better navigate the automated first phase of resume sorting.
Artificial intelligence tools such as the ones used on HiringSolved’s recruiting platform are also being employed more regularly to help reduce bias in recruiting. The company recently announced a new proof of concept tool called RAI (pronounced “ray”) that can help predict candidates’ gender and ethnic backgrounds to help companies reach diversity targets.
Shon Burton, the CEO of HiringSolved, explains that RAI takes a layered approach to candidate searches. Instead of applying filters that screen for viable candidates, the platform uses hundreds of data points to identify diverse candidates. It does this through a prediction engine that’s based on a proprietary statistical model the HiringSolved team developed in house. It allows users to “boost” search relevance by using the platform’s ethnic and/or gender diversity models. Those who meet certain criteria get pushed to the top of search rankings.
“We call it a diversity boost, because it’s not a filter, it’s actually changing the relevance algorithm,” he says. “The number of candidates doesn’t change,” adds Burton, “the relevant candidates will still be there, but I can turn on a boost for females so they bubble up to the top of the search.”
While these tools can help recruiters narrow down the millions of potential candidates into more manageable pools of more likely contenders, Burton suggests some may be taking it too far.
“There’s a flood of AI applications coming out, and a lot of them are going to be demoralizing to people,” he says. Burton chalks this latest development up to a pattern inherent in the way technology advances. “We gain the ability to do something, we do it, then we find out later that it wasn’t a great application of it,” he says.
Like restaurants that advertise “never-frozen” meat and produce, or telecommunications companies that advertise your ability to speak directly with a customer service representative, Burton believes there will soon come a time when resisting the conveniences of available technologies–many of which come at the expense of the end user–will become a major selling point.
“In 10 years it will be a brand differentiator,” he predicts. “I bet you’ll see companies saying, ‘You don’t have to talk to a machine to get hired here, because we believe in humans.'”
Burton adds that the real benefit of AI-based recruiting technologies lies in the initial screening phase, rather than during testing and interviewing phases. He describes HiringSolved as the Google for recruiters. With a price tag that starts at $5,000 per year, it allows them to search for positions in various industries by qualifications, experience, and location, as well as a variety of other filters, and then send an initial email to qualified candidates.
“Searching, refining the search, exporting the list, putting it into a campaign, creating a message and sending it, that’s up to three hours or more of work,” he said. “Our goal is to reduce the amount of time, give [the recruiters] back a few hours of their day, so they can focus on the human conversation.”
RAI is currently being piloted by some of the largest recruiting firms in the U.S., says Burton, and while most are eager to automate many aspects of the recruiting process, he hopes the human interaction element remains the same as it did before the invention of AI.
“Technology applied to recruiting is technology applied to one of the purest forms of human interaction,” he says. “Think about how horrifying it would be to talk to a machine about a pre-existing medical condition with someone in your family,” Burton observes.
While these technologies can help eliminate much of what Burton refers to as the ‘busy work’ of recruiting, he believes facial recognition, language analysis, and mood prediction tools take automation in recruiting a step too far.
“Where artificial intelligence is today in every industry, we’ve finally achieved the tipping point in just being able to do it, to build it, to make it useful,” he says. “The question now that everyone has to answer is, ‘Should we?'”