Strategy and Forecasting
The Strategy and Forecasting Track supports research on the long-horizon questions that advanced AI raises, from AI timelines and geopolitical competition to institutional futures in a post-AGI world. As capabilities accelerate, key decisions about advanced AI are being made with limited information, often before strong empirical evidence or broad consensus has emerged. Addressing these challenges requires structured forecasting, scenario analysis, geopolitical modeling, and macro-strategic thinking.
- Stage 1: Complete the general application, including 1-2 writing samples and track-specific short response questions
- Stage 2: Stream selection questions alongside possibly reference requests
- Stage 3: Interviews and work-tests
This track is focused on macro-strategy: understanding how the transition to advanced AI will unfold, how institutions and societies may adapt to highly capable AI systems, and what actions taken today can most improve outcomes in the long run. Some streams within this track center on structured forecasting, particularly quantitative work on capabilities, timelines, compute, and economic impact. Others emphasize scenario analysis and modeling, including frontier lab dynamics, geopolitical competition, state behavior, transition scenarios, and tabletop exercises.
These forecasts and models may be used to support analysis of what policy options are available to steer the path to AGI or the post-AGI future, to describe the costs and benefits of these options, and to raise awareness of how choices being made today could expand or narrow the range of options available to future policymakers. Research in this track might explore how, why, when, and where advanced AI will prompt rapid changes in industrial production, military tactics, and general scientific research, as well as the second-order effect of these changes on geopolitics, democracy, and capitalism.
Fellows in this track need to be comfortable with uncertain inference, probabilistic claims, and writing clearly about questions where the evidence base is limited. Experience with or interest in interdisciplinary research is helpful, as many of the research questions in this track ask how changes in one area of society will affect behavior in other fields. Strong candidates from past cohorts have come from forecasting, economics, history, philosophy, political science, international relations, computer science, security studies, and quantitative social science, among other backgrounds.
Fellows are matched to mentors based on fit, and projects are scoped to produce concrete artifacts (e.g., forecasting reports, scenarios, policy memos, strategic analyses, and peer-reviewed research) by the end of the program. Target audiences for the work produced in this track include lab strategy and policy teams; AISI staff; national security analysts; the funders and policymakers making long-horizon decisions about advanced AI; and the broader forecasting, governance, and AI safety communities.
We are interested in mentoring projects in AI forecasting, governance, and strategy. We want to improve the epistemic environment around thinking about AI risk by putting out concrete scenarios, exploring different dynamics and worlds that might happen, trying to forecast how things will go, and sometimes talking about how to deal with the risks. We hope that having better resources for thinking about the future will lead to better outcomes.
We aim to generalize tools for analyzing the dynamics of large-scale agency and power, such as public choice theory, to the setting in which machine minds are competitive with humans.
This stream focuses on how advanced AI could enable new and dangerous bio technologies, and on assessing when risks become tractable or urgent as those capabilities arrive.
My stream focuses on preserving checks and balances as governments adopt increasingly powerful AI. Fellows will work on questions like how Congress can maintain oversight of an AI-accelerated executive branch (including via privacy-preserving AI auditors) and what a positive vision for government AI adoption looks like. Projects will typically produce a public report and sometimes involve engaging directly with policymakers and other stakeholders.
This stream focuses on identifying tractable policy and technical interventions to gradual disempowerment, focusing on economic disempowerment and the intelligence curse. Possible project areas include:
- Formalizing intelligence curse dynamics into a model that can be tracked and monitored.
- More durable policy solutions to mass unemployment than UBI.
- Technical interventions to extend the centaur period.
My stream aims to prevent catastrophic outcomes through the AI transition by focusing on three lines of effort: nuclear and AI-enabled strategic threats; cognitive security, information competition, and AI-enabled influence; and moral game theory, competitive strategy, and peace. Fellows can pursue their own research, help build the field, or support grantmaking in any of these areas.
This stream works on infrastructure for AI safety research: AI tools that give safety researchers uplift, mechanism design and product development for funder coordination, and AI policy scenarios and proposals.
This stream focuses on forecasting, real world applications of world modeling with LLMs, formal & semiformal verification, capability evaluations design, coordination mechanisms, collective intelligence applications, and gradual disempowerment.