AI Safety Grant Program: Funding for High-Impact Projects

About the Grant Program

This grant program supports projects that aim to reduce existential risks from AI across four underexplored areas:

This initiative provides $4.5 – 5.5M USD annually to support projects that strengthen human capabilities and cooperation architectures in AI safety.

Funding Areas

1. Automating Research and Forecasting

Reference materials: Superhuman Automated Forecasting, Decision Forecasting AI & Futarchy, Superhuman Scientific Literature Research.

2. Neurotech for Human-AI Integration

Reference materials: Foresight Institute Neurotech Reports, Whole Brain Emulations, A Hybrid Approach to AI Safety.

3. Security Technologies for AI Systems

Reference materials: AI Infosec, Cryptographic Technologies, Security Without Dystopia.

4. Safe Multi-Agent Scenarios

Reference materials: Foundations of Cooperative AI, Multi-Agent Safety Hackathon, Paretotopian Goal Alignment.

Application and Evaluation Process

How We Evaluate Projects

Applications are accepted year-round and reviewed quarterly, with deadlines on the last day of March, June, September, and December.

Application Process

  1. Fill out the grant application form.
  2. Review by at least three technical advisors.
  3. If shortlisted, attend a brief interview.
  4. Decision made within 8 weeks after the deadline.

Additional Details

For detailed information about the grant program, including eligibility, focus areas, and application guidelines, please refer the document below: