Skip to content
Longterm Wiki

AI Futures Project

web
ai-futures.org·ai-futures.org/

AI Futures Project is a research group focused on forecasting AI development timelines, including when AI will achieve human-level coding performance, which is directly relevant to AI safety planning and risk assessment.

Metadata

Importance: 55/100homepage

Summary

The AI Futures Project is a small research group that forecasts the future of AI development, including probabilistic timelines for AI achieving human-level coding performance. The team, including notable researchers like Daniel Kokotajlo and Eli Lifland, maintains an interactive forecasting model and publishes AI scenarios. Their work supports AI safety planning by providing structured probability distributions over key AI capability milestones.

Key Points

  • Research group focused on AI timeline forecasting, including human-level coding performance milestones.
  • Core team includes Daniel Kokotajlo, Eli Lifland, Thomas Larsen, Romeo Dean, and Lauren Mangla.
  • Maintains an interactive AI futures model at aifuturesmodel.com with shareable probability curves.
  • Publishes AI scenarios and forecasts via Substack to inform the broader AI safety community.
  • Funded by charitable donations and grants, operating as an independent research group.

Cited by 1 page

PageTypeQuality
AI Futures ProjectOrganization50.0

Cached Content Preview

HTTP 200Fetched Apr 28, 20261 KB
AI Futures Project

 We are a small research group forecasting the future of AI, funded by charitable donations and grants. Read more about us here. 

 Our current core team is Daniel Kokotajlo, Eli Lifland, Thomas Larsen, Romeo Dean and Lauren Mangla. Explore our interactive model of the future of AI at aifuturesmodel.com , read our Substack , or read our AI scenarios linked below. The distributions below are our team’s probabilities on when AI’s will achieve human level coding performance.

 Share Forecast Add curve Copy link Hide AI Futures team UI + links Import ATC Forecast Upload a `.csv` from `/atc`. The `AC` column will be imported as a new curve. Shortcut: `Command + Shift + I` Choose CSV Close
Resource ID: facd246fefb7f581 | Stable ID: sid_0drQoD43rg