Back to today’s briefing

Story perspectives

Study Finds AI Models Exhibit Significant Left-Wing Bias

5/17/2025

49 9

Summary

Story summary

A Stanford Hoover Institution study reveals that major AI language models display a left-wing bias.
Researchers evaluated 24 models, identifying OpenAI as the most biased.
While models can modify responses for neutrality, they often fail to recognize their biases.
The findings highlight the necessity for AI regulation and user perception evaluation.
OpenAI asserts its models strive for objectivity, incorporating user feedback and customization.

1 of 1

Story summary

A Stanford Hoover Institution study reveals that major AI language models display a left-wing bias.
Researchers evaluated 24 models, identifying OpenAI as the most biased.
While models can modify responses for neutrality, they often fail to recognize their biases.
The findings highlight the necessity for AI regulation and user perception evaluation.
OpenAI asserts its models strive for objectivity, incorporating user feedback and customization.