<![CDATA[Educational transformation in the Society 5.0 era requires the implementation of the Merdeka Curriculum supported by Artificial Intelligence (AI) to strengthen 21st-century competencies. Although teachers demonstrate positive attitudes toward AI, its use remains largely limited to administrative tasks rather than supporting adaptive, data-driven learning. This study aimed to conduct a needs assessment for developing the Adaptive Merdeka Learning (AML) model by integrating the Merdeka Curriculum, Human-Centered AI, Adaptive Learning, Learning Analytics, and Competency-Based Education. The study represents the first phase of Design-Based Research (DBR) using a descriptive quantitative approach. The research was conducted in vocational high schools in West Sumatra, Indonesia, selected purposively based on their implementation of the Merdeka Curriculum, AI readiness, differentiated learning practices, and learning analytics capacity. Participants included school principals, teachers, and students. Data were collected through questionnaires, observations, semi-structured interviews, and document analysis. Instrument validity and reliability were established using Aiken's V and Cronbach's Alpha, respectively. The findings indicate that teachers have a good understanding of the Merdeka Curriculum and positive perceptions of AI; however, challenges remain in implementing differentiated instruction, AI-supported diagnostic assessment, personalized learning, and learning analytics. These findings reveal a gap between current instructional practices and the pedagogical demands of the Merdeka Curriculum. The study concludes that the development of the Adaptive Merdeka Learning (AML) model is essential to support learner-centered, AI-enhanced education and strengthen 21st-century competencies. ]]>
