Study Shows AI Models Still Far Behind Humans in Visual Creative Tasks, Scoring Only 59% vs Humans' 89%

This is a Plain English Papers summary of a research paper called Study Shows AI Models Still Far Behind Humans in Visual Creative Tasks, Scoring Only 59% vs Humans' 89%. If you like these kinds of analysis, you should join AImodels.fyi or follow us on Twitter.

Overview

• Creation-MMBench evaluates creative intelligence in multimodal language models
• Focuses on context-aware creativity using visual stimuli
• Tests four creative abilities: divergent thinking, convergent thinking, elaboration, and adaptation
• Includes 820 high-quality questions across five domains
• Reveals significant performance gaps between AI models and humans
• GPT-4o achieves best model performance (59.44%), still far behind human experts (89.18%)

Plain English Explanation

Creativity is a defining trait of human intelligence. While we've seen AI models get better at understanding images and text, measuring their creative abilities remains challenging, especially when they need to respond creatively to visual content.

The researchers built [Creat...

Click here to read the full summary of this paper