The opportunity
OpenAI’s mission is to ensure that general-purpose artificial intelligence benefits all of humanity. We believe that achieving our goal requires effective engagement with public policy stakeholders and the broader community impacted by AI.
What you'll do
Lead the development of applied AI risk standards. This includes working with: technical and policy teams to define and advance standards for mental health, youth safety, age assurance, efficacy, reliability, and privacy-related risks, with a particular focus on evaluations and assessments.
Work with technical owners to turn research, internal policies, and: evaluation methods into credible test methods, assessment criteria, and standards grounded in real system behavior.
Bring emerging requirements into internal planning and work with responsible: teams to clarify their implications for evaluations, controls, and assurance for our products.
Work with technical and GRC partners on assessor competence, independence,: conflicts of interest, evidence access, reporting, and the distinction between management-system audits and technical evaluations.
Contribute drafting, representation, and coordination on frontier: risk-management and independent-assessment standards.
Lead assigned standards-body workstreams, negotiate proposals, and coordinate: internal input, reducing the external engagement burden on technical experts. Contribute to relevant ISO/IEC, INCITS, NIST consortium, and cross-industry work.
What they're looking for
- Maintain clear approvals, negotiating positions, contribution records, and: implementation handoffs so that external commitments remain connected to technical owners and available capacity.
- Have technical depth in evaluation science, AI evaluations and red teaming,: AI safety, or trust and safety, combined with demonstrated ability to translate that expertise into credible standards or assessment methods.
- Have authored or materially shaped standards, evaluation criteria, control: frameworks, or normative technical contributions—not only coordinated participation.
- Can translate a broad safety objective into a clear requirement and explain: what evidence would demonstrate that it has been met.
