ISO/IEC JTC 1/SC 42 is the international subcommittee that develops standards for artificial intelligence. Its work covers core AI terminology, risk concepts, trustworthiness, and guidance that can be used to compare systems across vendors and countries. Because AI standards often shape how products are evaluated, this committee helps define the language of secure and reliable AI.
In cyber security, that matters because standards can become the baseline for audits, procurement, logging, incident response, and model governance. If a standard defines what counts as trustworthy behavior or adequate testing, it can influence how defenders assess model outputs, data handling, and failure modes. Attackers also benefit when organizations lack common controls, since inconsistent terminology and weak verification make it easier to hide model abuse, prompt injection, data poisoning effects, or unsafe automation. SC 42 does not enforce security itself, but it helps set the technical vocabulary that later becomes policy, compliance, and defensive practice.