This study presents the development and evaluation of a system designed to provide real-time feedback on graphic design work using foundation models. Aimed at enhancing human design capabilities with artificial intelligence, the system offers constructive, actionable feedback based on predefined design criteria. A total of 325 real-world product advertisement posters were evaluated, with the system demonstrating a high accuracy rate of 97.23% in classifying irrelevant images and delivering specific, actionable feedback in four key areas: color theme, font selection, layout level, and visual importance. The findings highlight the potential of AI to complement traditional teaching methods by offering personalized and timely feedback to support student learning. The quality of the feedback was further validated by instructors, who rated its usefulness at 4.2 out of 5. A cost analysis revealed the system to be economical, with an average cost of $0.007 per image, making it well-suited for large-scale educational use. Additionally, this study explores the ethical considerations of using AI-driven feedback systems in education. The findings underscore the potential of AI-driven feedback systems to enhance traditional teaching methods by offering personalized and timely feedback to support student learning.