Home

Ottawa · Dual US and Canadian citizen · Remote

Jeremiah Stanghini

0
0.5
1
Jeremiah Stanghini

I forecast questions that can’t be answered with certainty, and I build the operational machinery that lets small organizations act on answers like them.

Forecasting work since 2011, with organizations including Good Judgment Inc., Metaculus, the Swift Centre, RAND, the Forecasting Research Institute, and Mercor. First of 350 in the 2015 Good Judgment Project, and top-ten finishes in the Good Judgment, Metaculus, and RAND rankings in the years since.

Most of that work now points at artificial intelligence: serving as part of the human benchmark that AI forecasting systems are measured against, quantifying biosecurity risk from large language models, and modelling long-horizon AI risk scenarios for policy audiences.

Since 2025, operations at the Future of Life Foundation: fiscal sponsorship, grant and fellowship operations, contractor and vendor systems, and cross-entity compliance across two nonprofit entities.

From 2015 to 2024, digital transformation in the Government of Canada, as one of Canada’s Free Agents: AI product management, digital identity and credentials, data governance, and change management.

On the record
2009–2016
Authored six hundred articles on decision-making, psychology, and strategy
2011–2021
Affiliate faculty, strategy and decision-making, George Mason University
2015–2024
Government of Canada: Transport Canada, Canada Revenue Agency, Fisheries and Oceans, Canadian Coast Guard, Elections Canada, Canada’s Free Agents
2015
Certified Superforecaster, Good Judgment Project
2019
National Creativity and Transformation Award, Transport Canada
2025–
Operations Manager, Future of Life Foundation
Ongoing
Full history on LinkedIn
Selected forecasting on AI
2025
Longitudinal Expert AI Panel. One of sixty forecasting monthly on AI capabilities, adoption, and impact
2025
Forecasting biosecurity risk from large language models. One of twenty-two quantifying AI-enabled epidemic risk
2024
AI forecasting capabilities. One of thirty-nine serving as the human benchmark
2023
Adversarial collaboration on AI risk. One of ten analyzing AI safety scenarios
2023
Conditional trees. One of eight modelling long-horizon AI risk
2022
Existential Risk Persuasion Tournament. One of eighty-nine evaluating global catastrophic threats
Writing
2011–2019
A personal blog on decision-making, psychology, management, and whatever else I was reading at the time. Written under a different frame than the one I use now, and kept online as written. Read the archive.
Correspondence
LINKEDIN  ·  UPDATED SEPTEMBER 2026