Research Scientist, FSF Risk Modeling and Governance, DeepMind

We are hiring a research scientist to lead governance research and risk modeling work for the Frontier Safety Framework (FSF).

The FSF is Google’s framework for addressing possible severe risks from its frontier AI models. Both risk modeling and governance innovation are key to its continued success. Risk modeling helps Google understand the extent of the severe threats our models may pose, what mitigations are appropriate for addressing them, and whether, after the mitigations are applied, risk has been brought to a low enough level for different kinds of deployment. Frontier AI risk assessment is still nascent, and we are strengthening our efforts here in preparation for AGI.

Governance innovation is critical given the arrival of new risks on the AI frontier, as well as the shifting technical and geopolitical environment. Safety frameworks like the FSF have changed substantially since their conception, and must continue to evolve to confront the challenges ahead. Working with our policy and governance teams to broaden their reach and strengthen their application will also be crucial to their long-term success.

You will lead a team of researchers to advance both of these priorities. Practically, you will drive research and development of the FSF, lead qualitative and quantitative modeling of possible AI severe risks, and represent the FSF risk assessment program in internal and external governance (e.g. in model launch processes and the writing and publication of model safety reports).

Artificial intelligence will be one of humanity’s most transformative inventions. At Google DeepMind, we are a pioneering AI lab with exceptional interdisciplinary teams focused on advancing AI development to solve complex global challenges and accelerate high-quality product innovation for billions of users. We use our technologies for widespread public benefit and scientific discovery, ensuring safety and ethics are always our highest priority.

Responsibilities:

• Develop FSF research and write updates of the FSF to adapt to changing safety, governance, and policy opportunities and needs.
• Research and model pre-mitigation risks to ground our empirical risk assessment work (e.g. by defining FSF thresholds) and inform safeguard development.
• Work across evaluation and mitigation teams to assess the extent to which safeguards have reduced risk, and whether they are adequate for further development and deployment of frontier models.
• Manage a team of researchers and cross-functional collaborators working on projects, and represent the risk assessment program in internal and external governance.
• Write our external safety reports, and work with appropriate policy/governance teams to strengthen the frontier AI governance ecosystem. 

Minimum qualifications:

• PhD in Economics, Philosophy, or a STEM subject, or equivalent practical experience.
• 3 years of research experience in frontier AI safety or governance (e.g., in threat modeling, safety frameworks, or dangerous capability/alignment evaluations).

Preferred qualifications:

• Ability to communicate complex technical topics to leads, governance bodies and policy experts.
• Track record of driving complex projects, including in research settings.
• Track record of applying conceptual and quantitative reasoning methods to uncertain and under-constrained research or forecasting problems.