Researchers compared 22 language models and 102 human participants on three verbal creativity tasks to examine how AI influences creative thinking. Individual models produced original responses, but across models outputs were markedly more similar than human responses, indicating homogenization among AI-generated ideas. Adjusting randomness and prompts partially increased originality but did not resolve the uniformity, suggesting potential limits to AI-assisted brainstorming.
By contrasting human divergent and associative thinking with AI outputs on standard creativity tasks, the study highlights cognitive diversity, originality, and how training data shapes thought. This topic interests psychology because it probes how creativity emerges, the role of context and tools in thought, and the implications for human problem solving when using AI as a brainstorming aid.
Article Title: Scientists tested the creativity of AI models, and the results were surprisingly homogeneous
Link to PsyPost Article: https://www.psypost dot org/scientists-tested-the-creativity-of-ai-models-and-the-results-were-surprisingly-homogeneous/
Copy and paste broken link above into your browser and replace "dot" with "." for link to work. We have to do it this way to avoid displaying copyrighted images.
#AIcreative
#LanguageModels
#CreativityResearch
#DivergentThinking
#CognitiveFlexibility






