AUTOMATED EVALUATION OF HISTORICAL TEXTS: AN AI-BASED MODEL WITH STRUCTURED HISTORIOGRAPHIC CRITERIA

Abstract

This article proposes an artificial intelligence (AI)-based model, specifically a GPT assistant, designed to automate the evaluation of historical texts using structured historiographic criteria. Historical quality assessment frequently encounters challenges, such as cognitive biases and subjective inconsistencies. The proposed model incorporates seven evaluation criteria (disciplinary, epistemological, ethical, technical, pedagogical, reliability, and practical utility) aimed at ensuring a more objective, transparent, and consistent assessment. The article argues that a suitably trained GPT assistant can significantly mitigate common issues found in traditional manual evaluations, streamlining processes and improving overall consistency. The paper discusses potential advantages, acknowledges inherent limitations, and outlines avenues for future research, emphasizing the need for rigorous training data and critical historiographic supervision.