The process of ensuring that an AI system’s objectives and actions align with human values and intentions.