Translation: AI alignment is the process of designing, constraining, and guiding an AI system so that its behaviour and outputs remain consistent with human goals, values, and safety requirements. Its purpose is to minimise harmful or unintended behaviours, while recognising that no system can be made entirely risk‑free. Alignment ensures that an AI system operates within the ethical, operational, and safety standards established by humans.