AI bias is a major challenge, as machine learning models often inherit societal prejudices. While AI can be improved for fairness, it can't yet be fully trusted without oversight.