[The Editors' Verdict] How to Impress AI Interviewers

Lee Kyung-ho, Issue Team 2 Leader

For job seekers, ‘Gatmujeok’ and ‘Gatsanjik’ refer to Hyundai Motor Company. These terms mean office jobs and production jobs that even gods would envy. One of the important hurdles to joining Hyundai Motor is the artificial intelligence (AI) interview. According to one review, Hyundai Motor’s AI interview starts with a 1-minute self-introduction, followed by three sets of interview questions, each containing six questions, totaling 18 plus one question. Recent past questions range from specialized topics like explaining geometric tolerance to general ones such as “What would you do if everyone had to do a task they disliked?”, “Have you ever free-ridden without effort?”, and “If your manager asks you to finish a project but you have a planned vacation, what would you do?”

Coaching companies say, “The most important thing is that the applicant’s answers are analyzed based on Hyundai Motor’s internal evaluation criteria, so from the very first step of the 1-minute self-introduction to repeated questions and additional follow-up questions, the overall flow of answers must be consistent and trustworthy.” They also recommend finding experts with actual hiring experience as interviewers or HR managers. Many companies, including Hyundai Motor, have introduced AI competency tests and AI video interviews. For job seekers desperate for ‘quality jobs,’ it is essential to impress the AI interviewers along with having good qualifications.

Are there any problems with AI recruitment? When asked ChatGPT, it pointed out issues such as ▲algorithmic bias ▲data quality problems ▲lack of generalization ability ▲lack of transparency and interpretability ▲privacy concerns. Algorithmic bias means that since AI models learn based on training data, biases inherent in the training data can be reflected in the model, potentially causing prejudice against certain races, genders, ages, etc., leading to issues of diversity and fairness.

Hilke Shellman, an investigative journalist and assistant professor at New York University, revealed the flaws of AI recruitment in her recent book, “Algorithms.” She tested various recruitment software by posing as a job seeker. When she responded nonsensically in German, the AI rated her position highly. Another personality algorithm gave contradictory evaluations based on her various social media profiles. According to whistleblower reports received by Assistant Professor Shellman, people whose resumes included hobbies or special skills like ‘baseball’ and ‘basketball’ received some additional points. Conversely, those with the word ‘softball’ on their resumes were downgraded. In the U.S., people who write ‘baseball’ on their resumes are mostly men, while those who write ‘softball’ are mostly women. Assistant Professor Shellman exposed that such flawed systems can perpetuate discrimination and bias in the hiring process, causing serious harm and posing enormous potential legal risks for companies.

Despite numerous flaws and concerns, AI recruitment is spreading further, inevitably increasing its impact on job seekers, career changers, and current employees. AI has entered an era where it not only selects people but also evaluates current employees and even creates layoff lists. Although the final decision is made by humans, it has become a time when one must impress AI. Paradoxically, for AI recruitment to be properly established, more data must be accumulated and trial and error must be undergone. In that process, some will benefit while others will suffer. To reduce harm, one must think and act as AI desires. The recruitment dynamic between people is shifting to a dynamic between AI and AI.