Story perspectives

Study: AI Models Threaten Human Creativity and Diversity

3/13/2026

Summary

Story summary

Researchers at the University of Southern California warn that large language models like ChatGPT may diminish human creativity.
A review of over 130 papers finds LLMs produce less varied outputs than human thought, favoring dominant languages and ideologies.
This can cause users to internalize limited perspectives, affecting writing and reasoning.
The issue is compounded by policies of U.S. President Donald Trump’s administration that discourage AI models promoting diversity.

1 of 1

Story summary

Researchers at the University of Southern California warn that large language models like ChatGPT may diminish human creativity.
A review of over 130 papers finds LLMs produce less varied outputs than human thought, favoring dominant languages and ideologies.
This can cause users to internalize limited perspectives, affecting writing and reasoning.
The issue is compounded by policies of U.S. President Donald Trump’s administration that discourage AI models promoting diversity.