AI Offense-Defense Dynamics Researcher

Job SummaryAs AI capabilities rapidly advance, we face a fundamental knowledge gap: we don't yet fully understand the complex dynamics that determine whether AI systems, or even individual capabilities of them, predominantly threaten or protect society. In this role, you'll lead research to decode these offense-defense dynamics, examining how specific attributes of AI technologies influence their propensity to either enhance societal safety or amplify risks. You'll apply interdisciplinary methods to develop quantitative and qualitative frameworks that analyze how AI capabilities proliferate through society as either protective or harmful applications, producing actionable insights for developers, evaluators, standards bodies, and policymakers to anticipate and mitigate risks. This position offers a unique opportunity to shape how society evaluates and governs increasingly powerful AI systems, with direct impact on global efforts to maximize AI's benefits while minimizing risks. This role is 100% remote but requires occasional travel.
About CARMAThe Center for AI Risk Management & Alignment (CARMA) works to help society navigate the complex and potentially catastrophic risks arising from increasingly powerful AI systems. Our mission is specifically to lower the risks to humanity and the biosphere from transformative AI.We focus on grounding AI risk management in rigorous analysis, developing policy frameworks that squarely address AGI, advancing technical safety approaches, and fostering global perspectives on durable safety. Through these complementary approaches, CARMA aims to provide critical support to society for managing the outsized risks from advanced AI before they materialize.CARMA is a fiscally-sponsored project of Social & Environmental Entrepreneurs, Inc., a 501(c)(3) nonprofit public benefit corporation.
Responsibilities• Develop quantitative system dynamics models capturing the interrelationships between technological, social, and institutional factors that influence AI risk landscapes• Design detailed analytical models and simulations to identify critical leverage points where policy interventions could shift offense-defense balances toward safer outcomes• Expand and operationalize our current offense/defense dynamics taxonomy and nascent framework, developing metrics and models to predict whether specific AI system features favor offensive or defensive applications• Build empirically-informed analytical frameworks using documented cases of AI misuse and beneficial deployed uses to validate theoretical models• Research how specific technical characteristics (capabilities breadth/depth, accessibility, adaptability, etc.) interact with sociotechnical contexts to determine offense-defense balances• Build public understanding of offense-defense dynamics through blog posts, articles, conference talks, and media engagement• Create tools and methodologies to assess new AI models upon release for their likely offense-defense implications• Draft evidence-based guidance for AI governance that accounts for complex interdependencies between technological capabilities and deployment contexts• Translate research findings into actionable guidance for key stakeholders including policymakers, AI developers, security professionals, and standards organizations • A M.Sc. or higher in either Computer Science, Cybersecurity, Criminology, Security Studies, AI Policy, Risk Management, or a related field• Demonstrated experience with complex systems modeling, risk assessment methodologies, or security analysis• Strong understanding of dual-use technologies and the factors that influence whether capabilities favor offensive or defensive applications• Deep understanding of modern AI systems, including large language models, multimodal models, and autonomous agents, with ability to analyze their technical architectures and capability profiles• Experience in any of the following: Security mindset, Security studies research, Cybersecurity, Safety engineering, AI governance, Operational risk management, Systems dynamics modeling, Network theory, Complexity science, Adversarial analysis, or Technical standards development• Ability to develop both qualitative frameworks and quantitative models that capture sociotechnical interactions, and comfort creating semi-quantitative semi-empirical models also grounded in logic • Record of relevant publications or research contributions related to technology risk, governance, or security• Exceptional analytical thinking with ability to identify non-obvious path dependencies and feedback loops in complex systems • PhD in a relevant field• Experience with system dynamics modeling, hypergraph techniques, or other complex network analysis methods• Skills in developing interactive tools or dashboards for risk visualization and communication• Background in interdisciplinary research bridging technical and social science domains• Demonstrated aptitude in top-down techniques and first-principles thinking• Experience with the quantification of qualitative risk factors or developing proxy metrics for complex phenomena• Background in compiling and analyzing incident databases or case studies for pattern recognition• Familiarity with empirical approaches to technology assessment and impact prediction• Knowledge of international relations theory as it applies to technology proliferation dynamics