AI Shaping Culture, Worldview More Than You Think, Researchers Warn

Study finds AI systems like ChatGPT reflect and reinforce a narrow 'WHELM' perspective, potentially homogenizing global communication and problem-solving.

Apr. 10, 2026 at 11:20am

Got story updates? Submit your updates here. ›

A highly detailed, glowing 3D macro illustration of a complex circuit board or data server infrastructure, with neon cyan and magenta lights illuminating the intricate hardware components, conceptually representing the powerful yet unseen digital forces shaping modern culture and communication.As AI systems become ubiquitous, their influence on global culture and communication raises concerns about the potential loss of diverse perspectives and problem-solving approaches.Los Angeles Today

Researchers at the University of Southern California have found that the widespread use of large language models (LLMs) like ChatGPT may be subtly shaping human culture and communication in concerning ways. Their study reveals that these AI systems often reflect a Western, high-income, educated, liberal, and male (WHELM) perspective, which could gradually narrow the diversity of ideas, traditions, and problem-solving approaches that people are exposed to globally.

Why it matters

Cultural diversity is crucial for fostering creative solutions and a rich understanding of the world. If AI systems continue to promote a narrow set of values and communication styles, it could lead to a troubling 'cultural homogenization' that diminishes this important source of human knowledge and problem-solving capacity.

The details

The researchers, led by Yalda Daryani and Morteza Dehghani at USC Dornsife, reviewed studies across psychology, computer science, and linguistics to understand how LLMs perform across cultures. They found that these systems often favor Western values like individual freedom and fairness over ideas like tradition, authority, and community that are more central in non-Western cultures. This pattern also extends to how the AI systems handle subtle social interactions, with users' communication styles gradually converging toward a narrow WHELM norm.

  • The research was published in April 2026 in the journal Policy Insights from the Behavioral and Brain Sciences.
  • ChatGPT, one of the most prominent LLMs, debuted publicly in 2022 and has since seen a surge in usage, with hundreds of millions of people using such tools weekly.

The players

Yalda Daryani

A Ph.D. student in social psychology at the USC Dornsife College of Letters, Arts and Sciences and the lead author of the study.

Morteza Dehghani

A professor of psychology and computer science at USC Dornsife and the head of the Morality and Language Lab, who provided guidance for the research.

Zhivar Sourati

A Ph.D. student at the USC Viterbi School of Engineering and a co-author of the report.

Got photos? Submit your photos here. ›

What they’re saying

“AI isn't just reflecting culture anymore. It's actively shaping it. It's deciding what sounds polite, what sounds clear, even what counts as a good answer.”

— Yalda Daryani, Ph.D. student in social psychology at USC Dornsife

“When you ask AI for advice, you're not getting a neutral answer. You're getting the perspective of a very specific group of people, even if it doesn't say that explicitly.”

— Yalda Daryani, Ph.D. student in social psychology at USC Dornsife

“The more we rely on these systems, the more their outputs become part of our shared knowledge, and then that same material gets used to train the next generation of AI. So the cycle reinforces itself.”

— Morteza Dehghani, Professor of psychology and computer science at USC Dornsife

What’s next

The researchers outline a three-part approach to address the issue, including incorporating more diverse data sources during model training, consulting with experts from various cultural backgrounds, and evaluating model outputs based on multiple standards rather than a single Western-centric norm.

The takeaway

As AI systems like ChatGPT become ubiquitous, there is a growing concern that they may be subtly shaping global culture and communication in ways that diminish important sources of human knowledge and problem-solving capacity. Addressing this challenge will require concerted efforts to ensure AI development reflects the full diversity of human perspectives.