Color is the most influential visual factor that evokes emotional responses in humans. Nowadays, we can obtain color-emotion associations using various methods, including machine learning, deep learning, and fuzzy sets. This paper aims to identify to what extent these associations can be universal. We extend our previous work on color-emotion associations in art by examining these relationships in three additional datasets, each labeled with specific emotions. Utilizing fuzzy sets and logic approach, we extracted color-emotion associations from these new datasets and compared them using Jaccard and cosine similarity metrics to determine their resemblance to associations found in traditional art. Additionally, we conducted an experiment with AI-generated art images to compare color-emotion associations produced by artists versus artificial intelligence. Our results suggest that while most emotions exhibit universal color associations, emotions such as sadness, fear, shame, and trust display varying associations depending on the context. This variability could be attributed to these emotions being influenced less by color and more by the objects depicted within the art. Our findings highlight the complex interplay between color and emotional perception and suggest that while some color-emotion associations are universal, others are context-dependent. This study deepens our understanding of the emotional dimensions of color and broadens the applicability of these insights in fields like design and artificial intelligence.