<![CDATA[ABSTRACT This study aims to empirically examine the effect of the implementation of Artificial Intelligence (AI)-integrated Mobile-Assisted Language Learning (MALL) on classroom dynamics and students’ speaking proficiency in the Indonesian EFL context. The study is grounded in the persistent issues of low oral participation and high speaking anxiety in conventional classroom instruction. Addressing these challenges requires instructional approaches that promote interactive engagement while providing individualized learning support. An explanatory sequential mixed-method design was employed, involving 96 eleventh-grade students from the Explore Program at SMA NU 1 Gresik. Quantitative data were collected through pre- and post-speaking tests assessed using an analytic rubric covering fluency, accuracy, pronunciation, and vocabulary, while qualitative data were gathered through semi-structured interviews and classroom observations. The experimental group participated in AI-supported speaking activities for eight weeks, whereas the control group received conventional instruction with limited AI integration. The findings indicate that the experimental group demonstrated greater improvement in speaking scores compared to the control group, particularly in pronunciation and vocabulary development. Qualitative results reveal increased student engagement, more participatory classroom interaction patterns, a shift in the teacher’s role toward facilitation, and reduced speaking anxiety associated with individualized AI-generated feedback. These findings suggest that AI-supported learning environments may foster more responsive and interactive speaking instruction within classroom settings. Thus, the implementation of AI-integrated MALL contributed positively to classroom dynamics and students’ speaking proficiency within the eight-week intervention period, without extending claims beyond the duration of the study. ABSTRAK Penelitian ini bertujuan untuk menganalisis secara empiris pengaruh penerapan Mobile-Assisted Language Learning (MALL) terintegrasi kecerdasan buatan (AI) terhadap dinamika kelas dan penguasaan berbicara siswa dalam konteks EFL di Indonesia. Latar belakang penelitian ini didasarkan pada rendahnya partisipasi lisan dan tingginya kecemasan berbicara siswa dalam pembelajaran konvensional. Studi ini menggunakan desain mixed-method eksplanatori berurutan dengan melibatkan 96 siswa kelas XI Program Kelas Explore di SMA NU 1 Gresik. Data kuantitatif diperoleh melalui tes berbicara sebelum dan sesudah perlakuan yang dinilai menggunakan rubrik analitik mencakup kefasihan, ketepatan, pengucapan, dan kosakata, sedangkan data kualitatif dikumpulkan melalui wawancara semi terstruktur dan observasi kelas. Kelompok eksperimen mengikuti pembelajaran berbicara berbasis AI selama delapan minggu, sementara kelompok kontrol menjalani pembelajaran konvensional dengan pemanfaatan AI yang terbatas. Hasil analisis menunjukkan bahwa kelompok eksperimen mengalami peningkatan skor berbicara yang lebih tinggi dibandingkan kelompok kontrol, terutama pada aspek pengucapan dan pengayaan kosakata. Temuan kualitatif memperlihatkan adanya peningkatan keterlibatan siswa, perubahan pola interaksi kelas yang lebih partisipatif, pergeseran peran guru sebagai fasilitator, serta penurunan kecemasan berbicara yang dikaitkan dengan umpan balik AI yang bersifat individual. Dengan demikian, penerapan MALL terintegrasi AI memberikan kontribusi positif terhadap dinamika kelas dan peningkatan penguasaan berbicara siswa dalam konteks intervensi delapan minggu, tanpa menggeneralisasi dampak jangka panjang di luar periode penelitian.]]>
