Who Is Biased? Me?
Artificial intelligence now influences how students are assessed, admitted, and supported, yet every algorithm carries a trace of its makers. Obed and Bright Boateng’s 2025 study finds that predictive systems often inherit social inequalities from their training data, rewarding patterns of privilege and mislabeling others as “at risk.” Automated essay scoring favors certain dialects, and admissions algorithms repeat historical bias. The authors argue that progress depends on human oversight, transparent audits, and inclusive design teams. Education must teach its machines to learn fairness before efficiency becomes their only lesson.
Full study in World Journal of Advanced Research and Reviews (2025)
Bigbossfarin, Bias and variance contributing to total error, 2021. Public domain (CC0 1.0).