298. Bias

Systematic errors in AI outputs due to prejudiced assumptions in the training data or algorithm, leading to unfair outcomes.