AI Trust and Safety Careers

Trust and safety spans work that identifies and reduces harms to users, platforms, or communities. Current OpenAI job postings illustrate several related paths. A User Safety and Risk Operations role describes triage and resolution of sensitive cases, policy/process work, and collaboration with Product, Engineering, Legal, and Policy. A Model Policy role focuses on investigating model failures and translating findings into behavioral policies, evaluations, monitoring, and safeguards. A Software Engineer, Scaled Abuse role describes detection, investigation, and enforcement systems, while a Trust and Safety Data Engineering role builds datasets and pipelines for abuse detection and safety measurement. These are specific teams and vacancies, not a universal taxonomy. Daily work can therefore look very different. Operations specialists may investigate cases and apply standards consistently; policy professionals may clarify rules and edge cases; analysts may measure trends and escalation outcomes; engineers build tools and detection systems; researchers study model behavior or safeguards. Some roles combine several areas. Read the posting for the target domain, decision authority, technical depth, collaboration, and casework expectations rather than assuming every role involves content moderation or requires machine-learning engineering. Useful skills vary with the path, but commonly include careful written reasoning, consistent judgment, policy interpretation, data literacy, cross-functional communication, and respect for privacy and due process. Some positions involve sensitive material or urgent escalations; employer postings can state those conditions explicitly. Trust and safety is adjacent to model safety, but platform abuse operations and technical safety research have distinct goals and methods. Candidates should use current role descriptions to find the work and preparation that fit their experience.