Transformasi Evaluasi Pembelajaran melalui Generative Artificial Intelligence dalam Penilaian Berbasis Kompetensi

Pebry Salsa Saputri (Sekolah Tinggi Agama Islam Al-Gazali Soppeng, Indonesia)
Uswatul Hasana (Sekolah Tinggi Agama Islam Al-Gazali Soppeng, Indonesia)
Abd Hapid (Sekolah Tinggi Agama Islam Al-Gazali Soppeng, Indonesia)
Fatahul Rahman (Sekolah Tinggi Agama Islam Al-Gazali Soppeng, Indonesia)
Risma Handayani (Sekolah Tinggi Agama Islam Al-Gazali Soppeng, Indonesia)

Al Kindi Jurnal Pendidikan Islam Multidisipliner
Vol. 2 No. 3 (2026): Agustus

The development of Generative Artificial Intelligence (Generative AI) has transformed the paradigm of learning evaluation through assessment automation, learning data analysis, and adaptive feedback. However, its implementation still faces various challenges related to assessment validity, ethical use of technology, algorithmic bias, and the importance of educators' roles in academic decision-making. This study aims to analyze the transformation of educators' roles in the Generative AI-based learning evaluation paradigm, assess the effectiveness of Generative AI in supporting formative assessment in competency-based learning, and formulate a framework for adapting Generative AI implementation that still upholds validity, ethics, and human control. The study used the Systematic Literature Review (SLR) method with a qualitative-analytical approach to scientific articles published in the last five years and obtained from reputable academic databases. Data were analyzed using thematic analysis through a process of identification, filtering, extraction, and synthesis of research findings. The results of the study indicate that Generative AI transforms the role of educators from assessment implementers to facilitators, evaluators, and pedagogical decision-makers utilizing AI as a support system. Generative AI has been proven to increase the effectiveness of formative assessment through rapid, objective, and personalized data analysis. However, its implementation still requires validation by educators to address issues of validity, bias, and ethics. This research produces an adaptive framework that positions Generative AI as a decision support system through integrated data validation, balanced evaluation of cognitive, psychomotor, and affective aspects, and the application of the human-in-the-loop principle in assessment decision-making. These findings are expected to serve as a conceptual reference in the development of an effective, adaptive, and ethical competency-based learning evaluation system.