Will Recruitment AI Ever Pass the Turing Test?

The Turing Test was developed by Alan Turing in 1950 to determine a machine’s ability to exhibit intelligent behavior equivalent to – or indistinguishable from – a human. The key question that fascinated Turing was, “Can machines think?”

There is much debate over whether the Turing Test has ever been passed by a machine. In 2014, a computer program called Eugene Goostman successful convinced judges appointed by the University of Reading that he was in fact a 13-year-old Ukrainian boy. In 2015, MIT researchers helped develop a system that passed a “visual” Turing Test, producing written characters that fooled humans. In 2016, Microsoft released Tay on Twitter, and this AI chatterbot, which some argued could pass a millennial Turing Test, starting tweeting racist things and had to be shut down. And the Hollywood movie Ex Machina – whose entire premise is based on the Turing Test – serves as a cultural reminder of how advanced machines have come.

In the recruitment industry, passing the Turing Test would be groundbreaking. What if interested people could learn about working with your company via artificial intelligence? What if artificial intelligence could conduct interviews, screen candidates, and even negotiate job offers? What if artificial intelligence could manage the end-to-end recruiting process?

Marketing AI vs. Functioning AI

Before going into whether AI is or will be prominent within recruiting technology, we first must consider what is AI and what it is not. I spoke with a few influencers in the recruiting technology industry, and they are hearing more bark than bite…

“The term artificial intelligence is being overused by lots of marketing folk at the moment, especially when applied to recruiting,” said Rob McIntosh, Principal Advisor to Intelligent Talent Advisors. “Sort of like we have seen with the term ‘big data’. True AI in recruiting is still a little way off, so we will have to get used to vendors and marketers calling chatbots and algorithms AI for a little while.”

“The term AI in recruiting is like employment branding 10 years ago,” said Jonathan Duarte, the co-Founder of Gohire, a chatbot development company. “We’ll hear it for a while, but very few vendors are actually doing it.”

“I think AI, in the right hands, can do wonders, helping recruiters to spend time on substantial tasks versus routine [tasks] and helping novices get up to speed,” said Irina Shamaeva, Partner and Chief Sourcer at Brain Gain Recruiting. “However, there’s lots of marketing noise in the space. It’s as if you if are creating new software for recruiters, you are supposed to say it does some machine learning, AI, or deep learning…. What happens underneath, in the code and data, varies.”

AI In Action

As the experts point out, there are some rather grandiose claims about AI in recruiting floating around, but there’s also some impressive innovation – even if it still sits at more of an automation level and less of an actual artificial intelligence level.

“I have already seen a few companies using smarter chatbot technology to take over the bulk of a 30+-person recruiting coordinator department,” explained McIntosh. “Scheduling interviews and providing the candidates with info and updates on status will be the first tasks going the way of a smarter algorithm.”

The chatbot referenced by McIntosh is not a singular occurrence. AI assistants, like Mya and  Olivia, are already claiming to optimize the recruiter’s time.

Similarly, MightyRecruiter’s vector-space matching algorithm uses machine learning to evaluate a resume much in the way a human would. The feature is able to go beyond ATS-style keyword matching to make actual inferences about a candidate based on overall experience and qualifications. It then serves recruiters the most relevant resumes for each of their job openings, which saves busy hiring professionals from having to unnecessarily evaluate unqualified or poorly qualified candidates.

This kind of elimination of remedial tasks is appealing, and AI clearly has the potential to do that in the recruiting space.

And the thinking is if computer systems are able to perform this kind of labor now, what does the future hold? At the moment, the potential for how Turing Test-verified AI could recruit is unknown. But how much longer will it be a point on the horizon rather than a near reality.