Model Policy Manager

About the Team The Safety Systems team is at the forefront of OpenAI's mission to build and deploy safe AGI, driving our commitment to AI safety and fostering a culture of trust and transparency. The Model Policy team aligns model behavior with desired human values and norms. We co-design policy with models and for models by driving rapid policy taxonomy iteration based on data and defining evaluation criteria for foundational models’ ability to reason about safety. Key focus areas include: catastrophic risk, mental health, teen safety and multimodal safety. About the Role Providing access to frontier AI systems raises complex questions around dual-use science and catastrophic risk. How should models respond to requests involving chemical synthesis, biological experimentation, or pathogen research? Where is the boundary between legitimate scientific inquiry and information that could enable misuse? How do we design policies that meaningfully reduce risk without unnecessarily restricting beneficial research? This is a senior role in which you’ll help shape policy creation and development at OpenAI for addressing biological and chemical risks. You will develop structured policy frameworks and taxonomies to guide safe model behavior. This role sits at the intersection of biosecurity expertise, AI safety research, and policy design. You will help ensure that frontier AI systems can support beneficial life sciences research, such as drug discovery, public health, and biosafety, while reducing the risk that these capabilities could be misused. This role is based in San Francisco, CA. We use a hybrid work model of 3 days in the office per week and offer relocation assistance to new employees. In this role, you’ll: - Design and maintain model policies governing chemical and biological risk, defining how models should safely handle dual-use scenarios. - Develop structured taxonomies of chemical and biological risk that inform model training data, evaluation benchmarks, and safety monitoring systems. - Translate biosecurity and chemical security expertise into actionable model behavior, working closely with research and engineering teams to operationalize policy in training and evaluation pipelines. - Develop a broad range of subject matter expertise while maintaining agility across topics. - Identify emerging risk vectors where frontier AI capabilities could meaningfully lower barriers to harmful activity and develop mitigation strategies. - Engage with internal and external subject-matter experts in biosecurity, biodefense, and chemical safety to ensure policies reflect real-world risk landscapes. You might thrive in this role if you: - Have strong domain expertise in chemistry, biology, biosecurity, or related fields and are motivated to translate that expertise into principled, operational policies that scale to Frontier AI systems. - Have experience researching LLMs, ML, AI, tech policy, moral reasoning, and/or enjoy classification problems. - Have extensive experience defining, refining and enforcing policies for ML models across training, evaluation, and deployment. - Are comfortable navigating ambiguous, high-stakes problem spaces, balancing risk reduction with the benefits of scientific openness and innovation. - Can reason about the benefits and risks of open-ended problem spaces, generate novel approaches under ambiguity, and take full ownership of end-to-end solutions from concept through execution. - Have experience working at the intersection of science, policy, or emerging technology, such as in life sciences research, national security, risk analysis, technology policy, or AI safety. Most relevant publications: - Introducing HealthBench - Preparing for future AI capabilities in biology - Safety evaluations hub - OpenAI GPT5 System Card - Evaluating Fairness in ChatGPT - Improving Model Safety Behavior with Rule-Based Rewards - OpenAI Model Spec About OpenAI OpenAI is an AI research and deployment company dedicated to ensuring that general-purpose artificial intelligence benefits all of humanity. We push the boundaries of the capabilities of AI systems and seek to safely deploy them to the world through our products. AI is an extremely powerful tool that must be created with safety and human needs at its core, and to achieve our mission, we must encompass and value the many different perspectives, voices, and experiences that form the full spectrum of humanity. We are an equal opportunity employer, and we do not discriminate on the basis of race, religion, color, national origin, sex, sexual orientation, age, veteran status, disability, genetic information, or other applicable legally protected characteristic. For additional information, please see OpenAI’s Affirmative Action and Equal Employment Opportunity Policy Statement. Background checks for applicants will be administered in accordance with applicable law, and qualified applicants with arrest or conviction records will b

Strong judgment about how advanced AI systems may affect real-world risk, especially in ambiguous, fast-moving, or high-impact areas. Experience building or applying policies, taxonomies, harm models, threat models, or risk frameworks for complex technical, social, or adversarial systems. Ability to move across domains without needing to be the deepest subject-matter expert in every area, while knowing when to seek expert input. Capacity to turn fuzzy questions into structured policy frameworks, evaluation criteria, operational guidance, and enforceable model behavior. Comfortable using empirical evidence, including evaluations, red-teaming results, deployment observations, and model failure modes, to inform policy decisions. Systems-thinking across policy, data, graders, classifiers, training, deployment safeguards, measurement, monitoring, and escalation workflows. Technical judgment about what model behavior can realistically be trained, measured, evaluated, and enforced at scale. Ability to work well across research, engineering, product, policy, domain experts, and operational teams. Strong written communication to articulate complex tradeoffs where safety, user value, and implementation constraints matter. A pragmatic approach to safety focused on reducing real-world risk while preserving legitimate, beneficial, and socially valuable uses of AI. Enjoyment of fast-paced, collaborative research environments where priorities shift as models, evidence, and risks change. Grounded in implementation details, empirical results, and what can actually be trained or measured.