EVALUASI PROGRAM PELATIHAN PERANCANGAN ASESMEN BERBASIS KECERDASAN ARTIFISIAL BERDASARKAN KUALITAS PRODUK ASESMEN GURU

Abstract

This study aimed to evaluate a training program on designing artificial intelligence (AI)-based assessments, focusing on the quality of the assessment products created by teachers. Although generative AI offers new opportunities for assessment design, teachers' ability to utilize it pedagogically requires strengthening through structured training. An evaluative research method with a descriptive quantitative approach was applied to 25 teachers participating in a five-stage training program: content delivery, AI demonstration, a design workshop, independent product development, and a presentation and feedback session. The products were assessed using a five-aspect rubric—alignment with learning objectives, instrument quality, authenticity, AI integration, and rubric quality—on a scale of 1 to 4. The results showed an overall average of 3.44 (categorized as "very good"), with the highest score in alignment with learning objectives (M=3.64) and the lowest in AI integration (M=3.08). 56% of participants achieved the "very good" category, while 44% fell into the "good" category. The findings indicate that product-based evaluation provides an authentic picture of teachers' competency attainment while identifying areas requiring improvement, particularly regarding the critical and reflective use of AI.