The Korea Herald

지나쌤

[Newsmaker] Time to take the 'human' out of HR?

By Park Soong-joo

Published : Feb. 17, 2023 - 10:25

    • Link copied

(123rf) (123rf)

Looking to get hired? Your next job interview could be with a robot.

As artificial intelligence applications grow across industries, it is increasingly being used in human resources.

According to a survey conducted by Saramin, an online recruiting company, in 2022, nearly 2 out of 3 local companies said that AI recruiting programs have helped them save time and human labor.

At present, AI is mainly used by recruiters to save time and give more chances to more candidates by screening a lot more resumes and interviews. While AI programs will not replace humans altogether, utilizing the technology brings about the immense potential in human resources, experts say.

“(AI interviews) are a combination of pre-existing interviews and aptitude tests. It's not a substitute for the existing (interviews) but adds a new layer to them," said Seol Min-jun, who works as a recruiter and gives online lectures on AI interviews.

Needs for AI recruiting tools

The rise of AI recruitment tools has coincided with a shift in hiring trend in Korea. For years, group recruitment was the standard within the Korean job market, where job positions are mostly filled during a designated period. Now, more companies are turning to continuous recruitment, which means companies are always accepting resumes.

The shift in hiring pattern has led to human resource managers being overloaded with a barrage of resumes and job interviews.

To meet the need for remote interviews and to mitigate the workload of HR, more companies have been reaching out for AI tools like Viewinter HR, an AI interview tool created by Genesis lab. According to a survey conducted by Genesis lab, 91 percent of 204 human resources managers said that their companies have begun hiring constantly ever since the pandemic as group interviews have become unsafe, and that 37 percent of them said that they were considering installing AI tools.

How did AI interviews become the trend?

At anywhere at any given time, an interviewee can turn on the computer and start answering questions from the AI while making eye contact with the camera and speaking clearly to the microphone.

The process of assessing a candidate takes only minutes, with the AI program producing a specified assessment of the interview with a Behavior Even Interview score that measures both the quality of the answer and the soft skill presented by the interviewee.

Free from physical constraints like time and place, which in effect reduces the cost of recruitment, AI interviews could be conducted with more job candidates, allowing companies to consider more candidates.

“AI interviews might seem like added stress for job seekers, as they have one more thing to prepare for job hunting, but the AI could be the reason you get that job,” says Yuk Geun-sik from Genesis lab, which provides AI human resources programs including AI interviews for local companies.

Yuk Geun-sik, General Director of Genesis Lab's HR business division (Genesis Lab) Yuk Geun-sik, General Director of Genesis Lab's HR business division (Genesis Lab)

A current employee at Hyundai Motors, surnamed Oh, who had gone through a number of AI interviews before landing his job, recalls, “It seemed overbearing at first, but the more I prepared for it, I have come to prefer it over regular interviews.”

Another potential benefit, Yuk said, is that AI is able to assess the interviewees that come in first and the hundredth upon the same yardstick, without bias.

Can AI do you wrong?

Still, a lot of concerns and criticisms regarding AI recruiting remain.

Speaking from recent experiences with AI interviews, a Kia Motors employee surnamed Lee questioned the credibility of AI’s assessment, criticizing its supposedly arbitrary criteria.

Another concern is that some AI tools make it difficult to explain their assessment, which then makes it impossible for companies to tell an individual why they didn’t qualify for the job.

In 2020, a civic group filed administrative litigation against public organizations that use AI recruiting tools for failing to provide job applicants with reasons for not hiring them. The court sided with the civic group, and the organizations have not been able to share the algorithm, for which they did not comprehend the criteria.

To this, Genesis lab stressed the importance of openness with the algorithm, that every assessment made by the AI should be explainable, and that AI should not be the only assessment tool.

“People mistake that AI makes the call for failing or passing the candidates, but that is not true. People in human resources may flip the decision for whatever reason, so it really is them who make the final say,” Yuk said, adding that all of the AI interviews are recorded, so recruiters can play back an interview.

An example of an AI interview assessment conducted with Viewinter HR (Genesis Lab) An example of an AI interview assessment conducted with Viewinter HR (Genesis Lab)

Can you learn to beat AI?

Then there is the issue of whether an AI can assess the truthfulness of an interviewee. “Systems are yet to be perfect enough to detect all cheating,” says Seol.

AI isn’t a lie detector, says Yuk from Genesis Lab. When the system detects foul play, it leaves a flag to which recruiters can go back to the recorded footage to determine.

Similarly, some raise concerns over interviewees sharing strategies to beat AI within communities.

Instructors and YouTubers that lecture on these strategies are growing popular, with some offering paid lectures.

The question is, do these strategies work? The answer is that it depends on the AI interview tool used by the company.

For programs that have a set of the same questions, it is possible to learn how to beat the program because there is only one answer key and criteria to target.

Yet AI programs that are customized for each company have different questions and criteria, meaning there is no way to guess the exact recipe.

Still, there seem to be technical limitations with AI interviews that cross over the question of credible assessment.

From Oh’s experience, brightening up the environment and speaking loudly helps interviewees boost their grades. “I was told articulating slowly and loudly will get anyone good scores on AI interviews. Once I took the advice, it all seemed so easy,” said Oh.

“Like any regular interview, practice makes results better for AI interviews. Start practicing with a camera and you will sound more confident. Strategizing with practice does not cross over ethics,” said Seol.

This captured image shows online lectures on AI interviews (Holix) This captured image shows online lectures on AI interviews (Holix)

Prospects of AI in HR

“The ability to assess and predict potential is the essence of AI, to which could be applied from sourcing out the most qualified candidate to the annual performance evaluation. Basically, all that requires making an assessment of people, AI could do it,” said Yuk.

With the same algorithm, the developers claim that AI can carry out performance evaluations, further educate employees, assess education progress, and even predict employees’ retirement.

Companies are able to selectively phase into available AI tools to meet their needs. They could choose to us AI to screen resumes, conduct interviews, check for plagiarism, monitor cheating, or even assess interviewers.

Seol also sees AI tools as irreplaceable in HR, especially in screening out resumes and CVs.

“We will soon see AI burgeoning its role in recruiting, and technology will become sophisticated enough to replace face-to-face interviews,” said Seol.

Yuk sees that in near future, AI’s role would expand in HR and that it would be used not only in recruiting but also in educating employees and general management.

“The fundamental function of an AI is to evaluate potential, so AI would be able to assist in any type of evaluation of people,” said Yuk.