Enhancing TPACK and Statistical Literacy through Generative AI–Based Adaptive Learning: A Mixed-Methods Study

Maryati, Iyam
Gumilar, Surya
Rahayu, Ayu Puji
Harun, Makmur

Mosharafa: Jurnal Pendidikan Matematika
Vol. 15 No. 1 (2026): January

Penelitian ini mengkaji dampak kerangka pembelajaran adaptif terintegrasi Generative Artificial Intelligence (GenAI; ChatGPT) terhadap peningkatan Technological Pedagogical and Content Knowledge (TPACK) dan literasi statistis calon guru matematika. Kerangka tersebut menerapkan interaksi dialogis berbasis mahasiswa, structured prompting, dan scaffolding dosen untuk mempersonalisasi eksplorasi statistika. Dengan desain kuasi-eksperimen mixed methods, penelitian melibatkan 72 mahasiswa (37 kelompok eksperimen dan 35 kontrol). Data kuantitatif dianalisis menggunakan uji t berpasangan dan ANCOVA, sedangkan data kualitatif dianalisis secara tematik. Hasil menunjukkan kedua kelompok meningkat secara signifikan, namun kelompok eksperimen memiliki skor akhir tersesuaikan yang lebih tinggi. Temuan kualitatif menegaskan peningkatan pemahaman konseptual, kemampuan desain pembelajaran berbasis teknologi, serta refleksi kritis terhadap etika penggunaan AI. Studi ini mendukung integrasi literasi AI dalam kurikulum pendidikan guru. This study examines the impact of a Generative Artificial Intelligence (GenAI; ChatGPT)–integrated adaptive learning framework on improving Technological Pedagogical and Content Knowledge (TPACK) and statistical literacy among prospective mathematics teachers. The framework employed student-driven dialogic interaction, structured prompting, and lecturer-guided scaffolding to personalize statistical exploration. Using a mixed-methods quasi-experimental design, 72 students participated (37 experimental, 35 control). Quantitative data from tests and questionnaires were analyzed using paired t-tests and ANCOVA, while interviews and observations underwent thematic analysis. Results showed significant gains in both groups, but the experimental group achieved higher adjusted posttest scores, indicating superior effectiveness of GenAI-integrated learning. Qualitative findings highlighted improved conceptual understanding, instructional design skills, and critical reflection on ethical AI use. The study supports embedding AI literacy and pedagogically grounded prompting within teacher-education curricula and institutional policy.