“The five classical main dimensions of trustworthy AI are generalizability, robustness, privacy, explainability, and fairness/biases.”