Responsible artificial intelligence in Indonesian undergraduate AI curricula based on a national document analysis

The rapid expansion of undergraduate artificial intelligence (AI) programmes has created an educational challenge. Graduates need strong technical preparation, and they must also be able to design and evaluate AI systems responsibly. This study examines how responsible AI and human-centred AI are publicly presented in undergraduate AI curriculum documents in Indonesian higher education. Using document-based curriculum analysis, we analysed 46 official public source records associated with 22 undergraduate or bachelor-equivalent AI programme entries and 30 AI-proximate course records from programmes with accessible course-level evidence. The coding framework measured technical depth, responsible AI integration, human-centred AI orientation, learning-outcome visibility, longitudinal integration and assessment alignment. Inter-rater reliability was substantial at the programme level, with Cohen’s kappa = 0.73, and strong at the course level, with kappa = 0.81. The findings show that Indonesian higher-education AI curricula are technically substantial but socio-ethically uneven. Programmes scored high on technical depth, with M = 16.0/24, but lower on responsible AI integration, with M = 3.4/18, and human-centred orientation, with M = 3.6/14. A rule-based typology identified four curriculum configurations. Technical-Ethical Integrator, 18.2%. Technically Strong, Ethics-Minimal, 31.8%. Technically Oriented, Partial Ethics, 31.8%. Emerging or Unverifiable, 18.2%. The study contributes a transferable curriculum-audit framework for science, technology, engineering and mathematics education and shows that responsible AI reform should move from isolated ethics content toward explicit learning outcomes, longitudinal integration and assessment-aligned capstone design.