The rapid spread of artificial intelligence and modern trends towards using its applications in assessment processes that have become directed towards authentic assessment have prompted researchers to pay attention to investigating the consistency of the results of the authentic assessment of experts as a criterion with the results of several other assessments (artificial intelligence, self, and peer assessment) and determining the assessment method that is most consistent with expert assessments. This research also aimed to study the differences between the results of the authentic assessment of students' graduation projects using different assessment methods. To examine this, the researchers used a rubric that was prepared in a way that suits the graduation project standards set by a group of specialists, and based on the descriptive quantitative approach with a cross-sectional design, where a group of graduation projects of students from the Faculty of Education at An-Najah National University in Palestine were evaluated as a random cluster sample by experts as a criterion and assessment using Chat-GPT, peer assessment, and self-assessment. The results showed that there was a strong consistency between the expert assessment and Chat-GPT assessment, with a Pearson correlation coefficient of R=0.897. Compared to the peer assessment, the results show no meaningful alignment with the expert assessment, where the correlation coefficient is R=0.380. Unlike the peer assessment, the results show weak alignment to the expert evaluation with R = 0.380. The self-assessment, in contrast, shows a moderate alignment with the expert assessment, with a Pearson correlation of R = 0.484. From the above, it can be inferred that, among the other methods, the AI-based assessment had the highest alignment with expert judgment.
