A check of AI predictions for accuracy, fairness, and possible harm.
It is like checking a referee after a game. A big win means little if one team got every bad call.
It checks AI guesses in hiring and loans. It finds wrong calls, unfair patterns, and missing proof.
AI-bias
An audit checks if some groups get unfair wrong calls.
Model Calibration
Calibration shows if the AI's stated chances are trustworthy.
XAI
XAI helps auditors trace the reasons behind a prediction.
AI-regulation
Audit results can support rule checks and accountability.