This working paper proposes ARISE (Academic Research Intelligence & Scholarly Evaluation), a three-stage system for academic quality assessment designed to address the structural shortcomings of citation-based metrics — including citation counts, h-index, journal impact factors, and quartile rankings in the era of artificial intelligence. The proliferation of generative AI has rendered traditional publication-based indicators increasingly unreliable, while the structural limitations of existing metrics, narrow citation networks, long publication cycles, the invisibility of multilingual scholarship, and the systematic penalisation of interdisciplinary work, have long disadvantaged the social sciences and humanities in particular. ARISE proposes a redesign rather than an abolition of peer review. The system operates through three independent stages: First, an AI evaluation in which a large language model trained on the academic corpus assesses the work along five quality dimensions. Second, an independent human reviewer evaluation conducted blind to the AI's report, using the same five dimensions. Third, a blind synthesis stage in which a third expert reconciles both reports and produces the final ARISE score. The five evaluation dimensions are Originality, Intellectual Rigor, Interdisciplinary Impact, Openness and Reproducibility, and Humanistic Dimension. The paper outlines the system architecture, the rationale for piloting in the social sciences (with museology and cultural heritage studies recommended as the initial sub-discipline), a four-phase roadmap toward implementation, and governance principles addressing bias, transparency, and resistance to gaming. The document is offered as a basis for inter-institutional dialogue with research assessment reform initiatives including CoARA, DORA, and UNESCO's Open Science framework, as well as with funding agencies and individual research institutions.
Mert Kofoglu (Tue,) studied this question.