Same Work, Different Feedback: Melissa Warr on Hidden Bias in AI

From Azadeh and Guy on AI · 24:47

Show by TechEDGE

Azadeh and Guy talk with Dr. Melissa Warr, assistant professor of educational design and learning technologies at New Mexico State University, about how AI can quietly reinforce inequity in classrooms. Melissa shares new findings comparing 2024 and 2026 models. AI scoring of identical student work has become less biased, but the written feedback has not. When the AI is told a student likes rap music instead of classical, it writes at a lower reading level, gives more direct instruction, and talks down to the student. This echoes schooling’s “hidden curriculum,” where some students are treated as less capable.The conversation covers why teachers should disclose when they use AI for feedback, and how generative AI’s social nature pushes TPACK toward a stronger focus on context. Melissa describes classroom activities that make bias visible. In one, students generated images of the Three Little Pigs set in India and Mexico and spotted the stereotypes. In another, veterinary science students found that AI recommended oxygen for a cat with pneumonia at a city clinic but not at a low-resource one. She also explains why she protects AI-free time each morning, and why “one-to-three” (students talking to each other while using AI) may matter more than one-to-one devices. Her activities are available at equityandai.net. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit guyonai.substack.com

More episodes of Azadeh and Guy on AI

More episodes

PodSearch is a free, independent podcast search engine. Read about PodSearch, the privacy policy or contact us.