Value context and insight. lkm@koreatimes.co.kr
Kookmin Univ. study maps gap between AI hype and bureaucratic reality
Professor Kim Do-hyoung of Kookmin University / Courtesy of Kookmin University
By Lee Kyung-min
As governments around the world move to integrate generative artificial intelligence (AI) into the machinery of public decision-making, a new study suggests that confidence in the technology may be lagging behind its promise.
A research team led by professor Kim Do-hyung of Kookmin University proposed a framework to measure what it describes as a critical disconnect: the gap between the actual maturity of AI systems and the expectations placed on them by stakeholders. The study, published in the peer-reviewed journal Technovation, argues that this divide could help determine whether the technology advances or falters in high-stakes public research and development programs.
Seeking a more systematic approach, Kim and his co-authors devised what they call the “Maturity-Expectation Gap,” or MEG, framework. The model draws on survey responses from experienced evaluators and pairs them with a machine learning analysis of academic research, aiming to measure both how ready the technology is and how much is expected of it.
Their findings point to a notable divergence.
Different groups — including evaluators, policymakers and researchers — often hold sharply divergent views of how capable generative AI systems are today. Where expectations outstrip perceived maturity, the study found, confidence in adopting the technology tends to erode.
“The larger the gap, the lower the willingness to rely on AI in decision-making,” Kim said. “Without managing that gap, implementation may trigger skepticism rather than confidence.”
The research also points to uneven prospects for adoption.
Some areas of evaluation — particularly those centered on structured data — appear more readily suited to generative AI. Others, especially those that depend on nuanced qualitative judgment, may require further advances in the technology or greater institutional readiness before they can be reliably put to use, according to the study.
This article was published with the assistance of a generative AI system and edited by The Korea Times.