Forecasting project
AI Treaty Momentum Index
An index for tracking political momentum toward international AI coordination.
With Oksana Kotelnikova
Oksana Kotelnikova and I built the AI Treaty Momentum Index for the Apart Research AI Forecasting Hackathon.
ATMI is a lightweight forecasting model for the near-term chance of a new international AI risk treaty or binding regulation. It encodes historical cases from climate, nuclear, biosafety, and finance using mechanism classes such as policy windows, capability shocks, verification feasibility, public salience, epistemic shifts, and geopolitical alignment.
The model supports both nowcasting from current governance signals and scenario-casting against possible future events. The project placed third in the hackathon in November 2025. It was built during a single weekend; its small sample, first-pass coding, and broad cross-domain comparisons are important limitations.
Suggested citation
Kotelnikova, Oksana, and Sofiia Lobanova. “AI Treaty Momentum Index (ATMI).” Apart Research AI Forecasting Hackathon, 2025.