How AI threatens national research assessment systems: An analysis of the Netherlands' SEP framework

Summary

This article examines how AI technologies (large language models and Agentic AI) threaten the future of national research assessment systems—large-scale, state-organized periodic evaluations of research quality and impact at universities and public research institutions. Using the Netherlands' Strategy Evaluation Protocol (SEP) as a case study, the authors argue that AI fundamentally challenges the assumptions underlying these assessment systems, particularly around trust, authenticity, and the ability to evaluate human-led research. The piece explores how different types of national assessment systems face varying threat levels from AI disruption.