Can AI Be Fairer Than Humans? [StoryPack-Genesislab ①]

“Can AI be fairer than people?” That is the question Digital Daily opens with in its serial feature on Genesislab. ViewinterHR, the company’s AI video-interview solution, is used in hiring at more than 150 large corporations and public institutions, including Hyundai Motor, LG Uplus, and the Korean Army. The first installment of the series goes behind that number to ask how the fairness was engineered.

ViewinterHR was trained on roughly 4 million real interview videos, each scored by professional interviewers, HR executives, and industrial psychologists. That volume is roughly equivalent to the lifetime experience of 220 professional interviewers. To verify evaluator consistency, the team had the same candidate scored repeatedly and dropped evaluators whose ratings did not align.

Anti-cheating measures include face-verified identity checks, cross-candidate answer-plagiarism detection, and proxy-taker blocks. To make the reasoning inspectable, the system highlights plus and minus segments inside the interview transcript so human reviewers can follow the AI’s trail. In Korea’s TTA AI trustworthiness evaluation, ViewinterHR passed all 13 applicable requirements and 41 verification items. In 2023 it took the Minister of Science and ICT award at the country’s inaugural “AI Trustworthiness & Quality” competition.

When customers were asked why they adopted the product, trustworthiness topped the list at 21 percent, with cost-efficiency second at 12 percent. Rather than claiming AI is categorically fairer than humans, Genesislab’s design leans into making fairness visible.

Source: Digital Daily (디지털데일리) — Behind ‘AI Interviewer’ Adopted by Korea’s Largest Firms [StoryPack-Genesislab ①]