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 Human-AI Interaction team aligns model behavior with desired human values and norms. We co-design policy with models and for models, and thus policies can be directly plugged into our safety systems. Key focus areas include defining success criteria for how we might evaluate foundational models’ ability to reason about safety, values, and questions of cultural norms, driving rapid policy taxonomy iteration, and addressing critical societal challenges like misinformation risk and how the model should respond in medical contexts.
About the Role
Providing access to powerful AI models introduces a host of challenging questions when it comes to content policy development and enforcement: How do we define model policies for what we do and don’t allow to be generated? To what end? How do we do this in such a way that is actionable, objective and sustains replicability?
As an early member of the team, you’ll help shape policy creation and development at OpenAI and make an impact by helping ensure that our groundbreaking technologies are truly used to benefit all people. The ideal candidate can identify and develop cohesive and thoughtful policies with a sense of urgency. They can balance internal and external input in making complex decisions, carefully think through trade-offs, and write principled, enforceable policies based on our values. This role is embedded in our research teams, focusing on training the model to make safe decisions itself, which is different from Platform Policy which governs policies for the use of OpenAI products.
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 model policies that govern safe model behavior in an objective and defensible way - e.g. how should the model respond in risky/unsafe scenarios? What does unsafe mean?
Develop taxonomies that inform data collection campaigns, as well as zero-shot classifier development. How do you construct a taxonomy that lay people, AI researchers and AI models understand and can immediately use?
Lead prioritization for safety efforts across the company for new model launches, understanding and addressing technical and reputational trade-offs.
Develop a broad range of subject matter expertise while maintaining agility across topics
Iterate policies based on feedback on enforceability and edge cases
Partner with internal and external researchers to adapt our taxonomies based on latest research and best practices
Experiment and design policy enforcement using our latest advancements in AI research
You might thrive in this role if you:
Have studied or have an interest in product policy, philosophy, linguistics, and moral reasoning, and/or enjoy classification problems.
Have experience defining, refining and enforcing content policies for immediate human review and ML model training
Can understand the operational challenges of enforcing policies and can incorporate this into policy design.
Can analyze the benefits and risks of open-ended problem spaces; can generate ideas required to solve ambiguous problems and take full ownership of the solution.
Are very familiar with policy and safety/responsibility questions related to AI specifically.
Relevant publications from this team:
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 do not discriminate on the basis of race, religion, national origin, gender, sexual orientation, age, veteran status, disability or any other legally protected status.
For US Based Candidates: Pursuant to the San Francisco Fair Chance Ordinance, we will consider qualified applicants with arrest and conviction records.
We are committed to providing reasonable accommodations to applicants with disabilities, and requests can be made via this link.
OpenAI Global Applicant Privacy Policy
At OpenAI, we believe artificial intelligence has the potential to help people solve immense global challenges, and we want the upside of AI to be widely shared. Join us in shaping the future of technology.
