Multidimensional Evaluation of Generative Artificial Intelligence: Perspectives from the Theory of Multiple Intelligences and the 4P’s Model of Creativity

This study examined generative artificial intelligence (GenAI) across emotional, cognitive, and creative domains based on the Theory of Multiple Intelligences and the 4P model of creativity. Three studies were conducted with university students in Shandong. Study 1 compared ERNIE Bot and Spark Desk in generating 16 basic and compound emotions, showing higher accuracy for basic emotions than compound emotions (p < 0.01); ERNIE Bot received higher ratings for sadness, surprise, fear, sadness–disgust, and joy–surprise (p < 0.05), though 17.4% of its images contained violent content. Study 2 found ERNIE Bot’s critical thinking comparable to middle school levels, while Spark Desk scored lower (p < 0.05). Study 3 showed student-created works were more original than AI-generated works (p < 0.05). Overall, GenAI exhibited uneven development across emotional, cognitive, and creative dimensions. Future research should develop multimodal evaluation frameworks and hybrid AI architectures to enhance comprehensive intelligence.