AI alignment is the effort to make AI systems behave in ways that are consistent with intended goals, human values, safety constraints, and operational requirements.
Alignment work includes model training, evaluation, policy design, guardrails, feedback loops, and governance practices that reduce the risk of harmful or unintended behavior.