Scaffolding Visual Analysis in CPAR with Human and AI-Generated Images

The rapid proliferation of AI-generated images has raised global concerns over authenticity and artistic authorship, impacting fields from media to education. This underscores the critical need for enhanced visual analysis skills, particularly within the Contemporary Philippine Arts from the Regions (CPAR) curriculum. Despite prior studies on traditional instruction, CPAR lacks structured frameworks to help students differentiate between human-made and AI-generated images, necessitating a systematic approach to develop analytical and creative skills. This study employs a quasi-experimental mixed-method design to examine the impact of a scaffolded Claim-Evidence-Reasoning (CER) framework compared to traditional instruction on students' ability to critically analyze AI-generated and human-created images. Findings indicate that students in the CER framework demonstrated significantly higher post-test improvements in visual analysis across Arts & Design (n=36, d=1.93), HUMSS (n=56, d=1.65), and ABM (n=116, d=0.9). However, variations in post-assessment scores suggest that the framework's effectiveness depends on trackspecific cognitive demands. While students in the experimental group reported that CER's structured approach deepened critical thinking and creative engagement, some encountered challenges in adapting to its rigid structure in disciplines requiring more fluid argumentation. This study confirms that structured scaffolding enhances students' analytical and creative abilities when engaging with contemporary visual culture. It highlights the need for track-specific instructional strategies and teacher training to optimize implementation. Future research should explore adaptive CER models integrating AI-driven assessment tools for personalized feedback, ensuring that scaffolded visual analysis strategies are effectively applied in CPAR and similar curricula.