<![CDATA[AI-based evaluation is increasingly applied in educational settings; however, research on junior high school students' perceptions of this technology particularly in the Indonesian context remains limited. Addressing this gap is critical, as student perceptions significantly shape the effectiveness of technology adoption in learning. This study aimed to analyze students' perceptions of AI-based evaluation and examine its influence on their learning process at SMP Kristen Satya Wacana Salatiga. A quantitative approach with a descriptive correlational design was employed. Data were collected through a 15-item Likert-scale questionnaire and observation sheets administered to 59 students in an Informatics subject. The instruments were validated and tested for reliability prior to use. Results indicated that students' perceptions of AI-based evaluation fell into the low category, although more than half of the students actively revised their answers after receiving system-generated feedback. Pearson correlation analysis revealed a significant positive relationship between student perceptions and the learning process, classified as moderate. These findings suggest that AI holds potential for supporting student engagement in learning; however, it cannot yet replace the teacher's role, particularly in providing contextual and dialogic feedback. ABSTRAK Evaluasi berbasis Artificial Intelligence (AI) semakin banyak diterapkan dalam pembelajaran, namun penelitian mengenai persepsi siswa jenjang SMP terhadap teknologi ini khususnya di Indonesia masih sangat terbatas. Kesenjangan ini penting untuk diatasi mengingat persepsi siswa berperan besar dalam menentukan efektivitas adopsi teknologi dalam pembelajaran. Penelitian ini bertujuan untuk menganalisis persepsi siswa terhadap evaluasi berbasis AI serta pengaruhnya terhadap proses belajar siswa di SMP Kristen Satya Wacana Salatiga. Penelitian menggunakan pendekatan kuantitatif dengan desain deskriptif korelasional. Data dikumpulkan melalui kuesioner 15 butir berskala Likert dan lembar observasi terhadap 59 siswa pada mata pelajaran Informatika. Instrumen telah diuji validitas dan reliabilitasnya sebelum digunakan. Hasil penelitian menunjukkan bahwa persepsi siswa terhadap evaluasi berbasis AI berada pada kategori rendah, meskipun lebih dari separuh siswa aktif memperbaiki jawaban setelah menerima umpan balik dari sistem. Analisis korelasi Pearson mengungkapkan adanya hubungan positif yang signifikan antara persepsi siswa dan proses belajar dalam kategori sedang. Temuan ini mengindikasikan bahwa AI berpotensi mendukung keterlibatan belajar siswa, namun belum dapat menggantikan peran guru, terutama dalam memberikan umpan balik yang kontekstual dan dialogis.]]>
