Program Manager, Responsible Scaling Policy

About the role

We are seeking a pragmatic generalist to help grow capacity for our ongoing efforts to execute on (as well as improve) Anthropic’s Responsible Scaling Policy (RSP). 

Our RSP is a series of technical and organizational protocols that we have adopted to help us manage the risks of developing increasingly capable AI systems. Among other things, it involves maintaining:

• Forward-looking plans for improving risk mitigation across the areas of Security, Alignment, and Safeguards. To maintain these plans, we work with stakeholders from across the company to set and iterate on ambitious but achievable goals, regularly check in on whether we’re on track to meet the goals, and problem-solve as needed to fulfill their spirit.
• Ongoing assessments of how well we are executing on the vision of our RSP —especially the extent to which our AI systems may or may not pose catastrophic risks, in light of their capabilities as well as our risk mitigations. We monitor developments in our AIs’ capabilities, and changes in our risk mitigations, to keep these up to date and ensure that any adjustments to our operational practices are made deliberately and thoughtfully.

As the company grows and AI capabilities improve, executing on our RSP is becoming increasingly important, complex, and cross-functional. Our RSP team works to:

• Set and check in on forward-looking goals for improving risk mitigations.
• Check in on activity across the company to ensure that our activities and risk mitigations are in line with the RSP.
• Partner with the Security, Alignment, and Safeguards teams to develop practical and effective risk mitigation strategies.
• Propose RSP revisions to reflect what we’re learning.
• Collaborate with our Policy team when the RSP is relevant to regulation.

This role offers a unique opportunity to work on the nuts and bolts of AI risk reduction, interacting with many different parts of the company. 

Success in this role means learning quickly about many aspects of our business; combining this with a deep understanding of the major catastrophic risks potentially posed by advanced AI; and bringing a pragmatic problem-solving attitude, constantly looking for the best and most practical ways to achieve risk reduction.

In this role, you will:

• Build relationships with internal experts on Safeguards, Security, Alignment, capability evaluations, and other relevant areas.
• Learn about ongoing questions and challenges faced by these teams, and help to find the best balance between AI risk reduction and pragmatism.
• Gather information from across the company and assist in analyzing relevant risks.
• Help to drive conversations about the ambitious-but-achievable plans we can make to level up our risk mitigations.
• Collaborate with our Policy team, advising on regulation relevant to RSPs.
• Generally work to identify the most promising risk mitigations to be further developed and/or exported to the rest of the AI industry.

You may be a good fit if you:

• Have (or can quickly develop) a deep understanding of major catastrophic risks potentially posed by advanced AI.
• Are passionate about reducing the harms and improving the benefits of AI  and align with Anthropic's safety mission and commitment to responsible AI development
• Are a pragmatic problem-solver who relishes the opportunity to find the best balance between the goal of reducing catastrophic risks and staying on the frontier of AI development.
• Can write quickly and clearly in a way that gives audiences an easy time understanding your reasoning and justifications for your claims
• Are a flexible generalist who is excited to learn about many topics and evolve your work in unanticipated direct