Science ❯ Computer Science ❯ Artificial Intelligence ❯ Ethics in AI
The revision elevates misalignment plus persuasion into formal risk thresholds with mandatory safety reviews before release.
