AI Policy Researcher, User Wellbeing, Trust and Safety

About the role

As an AI Policy Researcher on the Trust & Safety Responsible AI Policy team, you will champion the responsible development and deployment of frontier generative AI models. The focus is on model bias, political risk, and overall model behaviour to ensure safe, fair, and trustworthy AI experiences for users.

Key responsibilities

• Design and maintain multimodal GenAI policies covering user‑harm domains such as adult content, graphic material, consent, and mental‑health impacts.
• Translate risk and harm models into clear behavioural specifications, evaluation criteria, grading guidance, and system‑level safeguards.
• Define practical boundaries between beneficial AI uses and applications that could enable harm, exploitation, or unsafe outcomes.
• Partner with safety researchers, engineers, product teams, and other stakeholders to operationalise policy into scalable model behaviour and measurable safeguards.
• Develop end‑to‑end policy workflows from pre‑launch evaluation to post‑launch monitoring, including golden‑set construction, labeling guidance, calibration, adjudication, and coverage analysis.
• Leverage red‑teaming results, deployment data, and edge‑case analyses to continuously improve policy and evaluation quality.
• Identify emerging capability areas that may introduce new safety, fairness, or bias challenges.
• Monitor post‑launch model activity to spot gaps in the policy framework and update safeguards accordingly.
• Champion research to strengthen the defensibility of policy positions and incorporate external expert input.

Required profile

• Strong policy acumen with the ability to assess AI risks and propose actionable solutions.
• Excellent judgment, creativity, and analytical rigor.
• Proven ability to translate complex technical risk into clear guidance for diverse stakeholders.
• Experience working cross‑functionally with product, engineering, data science, legal, and public‑policy teams.

Required skills