Background: Empirical evidence indicates that junior high school students’ critical thinking skills in mathematics remain suboptimal, with national assessments highlighting gaps in analysis and evaluation competencies. Traditional teaching methods fail to address these higher-order thinking demands effectively. Aims: This study investigates the effectiveness of an integrated learning model combining Ethnomathematics, Augmented Reality (AR), Artificial Intelligence (AI), and Deep Learning (DL) in enhancing students’ critical thinking skills in mathematics. Method: A mixed-method sequential explanatory design was employed. The quantitative phase used a quasi-experimental one-group pretest-posttest design with 64 junior high school students in Purworejo, Indonesia. The qualitative phase included semi-structured interviews, focus group discussions, and classroom observations. Critical thinking was measured using a validated 10-item test (Cronbach’s α = 0.82). The intervention lasted four weeks (12 sessions of 80 minutes). Results: The quantitative phase showed significant improvement in critical thinking scores (t = 4.126, p < 0.001, 95% CI [18.2, 27.8]) with a high N-Gain of 73% (Cohen’s d = 0.58). Interpretation and evaluation indicators exhibited the highest gains (N-Gain 0.76–0.77). Qualitative findings revealed positive student engagement (mean 4.6/5) and cultural relevance (mean 4.5/5), while teachers found the model innovative but noted infrastructure and training challenges. Conclusion: The integrated Ethno-AR-AI-DL model effectively improves critical thinking and is positively perceived by users. These findings suggest the potential of culturally relevant, technology-enhanced mathematics instruction, provided teacher training and infrastructure are addressed.
Copyrights © 2026