Using the Uniqueness Quotient in the Assessment of Creativity of Image Generation Models

Keywords:
Generative AI, Objective evaluation, Visual originality, Multimodal generative AI models, Human-AI comparison

Abstract:
The rapid rise of Generative AI raises questions about its creative abilities, but most assessments focus on text or subjective visual reviews. This study uses the uniqueness quotient to compare the creativity of visuals generated by AI models (Intent-driven, Prompt-driven) and by humans (art and non-art students). Results show no significant difference in overall creativity between humans and AI. Subsample analysis finds Intent-driven models do not differ from art students in originality and outperform non-art students. Promptdriven models show unique originality patterns. Distribution analysis shows asymmetry: AI models keep a high average, but peak innovativeness remains unique to humans. This study's theoretical contribution is establishing the uniqueness quotient as a reliable, objective method for evaluating visual creativity. In practice, context-aware, intent-driven models can be integrated as advanced co-creators in creative industries, and the uniqueness quotient can serve as a measure of future machine creativity.

CITATION:

IEEE format

Copy to clipboard

M. Milošević, I. Ristić, M. Milošević, “Using the Uniqueness Quotient in the Assessment of Creativity of Image Generation Models,” in Sinteza 2026 - International Scientific Conference on Information Technology, Computer Science, and Data Science, Belgrade, Singidunum University, Serbia, 2026, pp. 240-246. doi:10.15308/Sinteza-2026-240-246

APA format

Copy to clipboard

Milošević, M., Ristić, I., Milošević, M. (2026). Using the Uniqueness Quotient in the Assessment of Creativity of Image Generation Models. Paper presented at Sinteza 2026 - International Scientific Conference on Information Technology, Computer Science, and Data Science. doi:10.15308/Sinteza-2026-240-246

BibTeX format
Download

RefWorks Tagged format
Download