AI Safety

The discipline of preventing AI systems from causing harm, whether through accidents, misuse, or unintended behavior. It spans technical robustness, alignment, and governance.