[ Round 1 Open ]

Forecasting AI Progress

A large-scale, comprehensive forecasting tournament dedicated to predicting advances in artificial intelligence.

Our aim is to significantly improve the accuracy of forecasts made on AI progress, enabling better preparation for its many anticipated impacts.

Join us as a forecaster and keep up with the results.

AI Forecasting Calibration Tutorial
Practice AI Questions
Win prizes by forecasting & elucidating the short, medium and long-term future of AI
$50K IN PRIZES
Bayes prize
$20,000
IN TOTAL PAYOUTS
Awarded to top 5 across all 90 short, medium and long-term forecasts.
Maximum Likelihood prize
$7,000
IN TOTAL PAYOUTS
Awarded to top 5 across 30 short-term forecasts.
Hill Climbing Prize
$9,000
IN TOTAL PAYOUTS
Awarded to top 5 across 30 medium-term forecasts.
Deep Learning prize
$7,500
IN TOTAL PAYOUTS
Awarded to top 5 across 30 long-term forecasts.
Dreyfus prize
$6,500
IN TOTAL PAYOUTS
Awarded to top 5 most insightful analyses of the future of AI.
See Full Tournament Rules & Prize Details
Outcomes

Our aim is to generate a massive dataset of artificial intelligence (AI) forecasts over a range of time-frames, train models to optimally aggregate those forecasts, and use these to build a highly accurate map of the future of AI.

Participants will be contributing to the development of important new insights about the future of AI. Our results will be made publicly available.

The tournament will provide a unique proof of concept for aggregate probabilistic forecasting in the realm of AI.

Advisory Board
Amanda Askell
OpenAI
Research Scientist on the policy team at OpenAI, where she’s worked on topics including responsible AI development, safety via debate, and publication norms in ML.
Joscha Bach
AI Foundation
Cognitive scientist and the VP of Research at the AI Foundation. Previously, AI researcher at MIT Media Lab and the Harvard Program for Evolutionary Dynamics.
Miles Brundage
OpenAI
Research Scientist on the Policy team at OpenAI and Research Affiliate at the University of Oxford's Future of Humanity Institute.
Owain Evans
Future of Humanity Institute
Research Scientist working on AI Safety at the Future of Humanity Institute, and member of the Board of Directors of Ought.
Saurabh Mishra
Stanford University
Manager of the AI Index Program at Stanford Institute for Human-Centered Artificial Intelligence (HAI).
Carroll ‘Max’ Wainwright
Metaculus
Research Scientist focused on technical aspects of AI safety, and Co-founder of Metaculus.