SIDLIT 2026

Artificial Intelligence (AI) is rapidly transforming higher education, yet many instructional implementations position students as passive consumers of AI-generated content rather than active participants in the learning process. Emerging research suggests that excessive reliance on AI tools may reduce learner autonomy, weaken self-regulated learning behaviors, and contribute to what scholars describe as “metacognitive laziness.” While AI-powered systems can improve academic performance and provide adaptive support, little research has examined instructional approaches that require students to critically engage with and improve AI-generated outputs.


This study proposes a human-AI co-design instructional model in which students actively evaluate, critique, and revise AI-generated content rather than simply accepting AI responses. Using a convergent mixed-methods research design, the study will investigate how this approach influences learner agency, cognitive load, and academic performance among undergraduate students in higher education. Approximately 40–60 students will participate in a quasi-experimental comparison between a traditional instructional environment and a human-AI co-design learning environment. Quantitative data will be collected through pre- and post-assessments, learner agency surveys, and cognitive load measurements, while qualitative data will be gathered through interviews, written reflections, and learning management system interaction logs. 


The study seeks to address three important gaps in current literature: the limited examination of students as active evaluators of AI outputs, the lack of understanding regarding cognitive load in AI co-design environments, and the unresolved tension between AI-driven personalization and learner autonomy. Findings are expected to provide practical guidance for educators, instructional designers, and higher education institutions seeking to integrate AI in ways that enhance critical thinking, learner engagement, and meaningful learning outcomes.