AIPhys.org Technologyabout 2 hours ago
Study Finds AI Bots Defy Stereotypes Harder
4 min read
A Rutgers University study published in Nature Machine Intelligence in 2026 reveals that large language models expend more reasoning tokens when processing counter-stereotypical information. Researchers Calvin K. Lai and Messi H.J. Lee tested five models, including OpenAI's o3-mini and Claude 3.7 Sonnet, finding that most echo human implicit biases during reasoning.
Level
Hype check
Tap to vote and see what everyone thinks.
