Ensuring AI systems behave according to human values and intentions, a critical area of AI safety research. Foundation of responsible AI development.