Skip to content
dotdock

Generative artificial intelligence models outperform students on divergent and convergent thinking assessments

2025

Publication cover

Open publication workspace · Sign in to read the full PDF.

AI-generated summary

1) Generative AI models now surpass human performance in both divergent and convergent thinking assessments, raising questions about current creativity evaluation methods.
2) This study compares the creative abilities of humans and three leading generative AI models (ChatGPT-40, DeepSeek-V3, and Gemini 2.0) using the Alternate Uses Task (AUT) and the Remote Associates Test (RAT).
3) The findings reveal that all tested AI models outperformed human participants in both originality and problem-solving tasks, with ChatGPT-40 demonstrating the highest performance.
Tags: Generative AI, creativity, divergent thinking, convergent thinking, artificial intelligence

Check the original publication for accuracy and context.