Develop and advocate a credible, pro-human AI strategy centered on reliable, verifiable, controllable, human-directed tools, while monitoring non-hyperscaling technical approaches such as smaller models, open-source AI, robotics, and industrial or scientific applications. Engage policymakers and other stakeholders, build coalitions around sovereign AI strategies, and translate the strategy into concrete policy recommendations.
The role involves conducting horizon-scanning for novel AI-enabled hazard classes and developing methodologies for reasoning about emerging threats. The researcher will also collaborate across programs to integrate civil safety perspectives with broader risk assessment and governance.
Develop and operate an evidentiary infrastructure combining ontologies, knowledge graphs, and LLM-assisted pipelines to systematize AI safety arguments. Architect agent coordination patterns and build export pipelines to translate structured argumentation into persuasive communications for various stakeholders.
