This study aims to categorize educational outcomes in generative AI-based art education research and to examine their distribution and emerging patterns. A systematic literature review was conducted on journal articles and theses published between 2017 and 2026, resulting in a final sample of 57 studies. Educational outcomes were coded into four dimensions: cognitive, affective, metacognitive, and instructional/ environmental. When multiple outcomes were reported within a single study, they were treated as multiple responses, and the analysis focused on the frequency and distribution of each category. The results indicate that educational outcomes are predominantly concentrated in the cognitive dimension, particularly in creativity, problem-solving, and expressive ability. In contrast, affective and metacognitive outcomes appear less frequently, and relatively few studies address instructional and environmental aspects. This distribution suggests that existing research tends to emphasize cognitive performance while giving comparatively less attention to emotional, strategic, and contextual dimensions of learning. This study contributes by providing a systematic categorization of educational outcomes and by identifying distributional tendencies across different dimensions. The findings highlight the need for a more balanced approach in future research, incorporating affective, metacognitive, and instructional/environmental perspectives alongside cognitive outcomes.