Investigating perceived fairness of AI prediction system for math learning: A mixed-methods study with college students

Yukyeong Song, Chenglu Li, Wanli Xing, Bailing Lyu, Wangda Zhu

Abstract: Highlights•Favorable prediction outcome is positively related to a higher perceived fairness of AI's decision.•Higher algorithmic bias and lower transparency are related to a lower perceived fairness of AI's decision.•Students who major in STEM and who have higher math anxiety tend to report a higher perceived fairness of AI's decision.•Accuracy of the prediction outcome is positively related to a higher perceived fairness of AI's decision.•Students desire transparency with guidance, explainability, interactive communication, constructive feedback, and emotional support.