We are interested in funding projects that address one or more of the following four areas:

Automating Research and Forecasting

  ● Scaling AI-enabled research to support safe AGI development
  ● Scaling efficient forecasting methods relevant for safe AGI
  ● Other approaches in this area

Neurotech to Integrate with or Compete Against AGI

  ● Brain-Computer Interfaces (BCIs) to enhance human cognition or facilitate
human-AGI collaboration
  ● Whole Brain Emulations (WBEs) as human-like intelligences that are interpretable and alignable
  ● Lo-fi emulations using behavioral and neural data with deep learning
  ● Other approaches in this area

Security Technologies for Securing (AI) Systems

  ● Implementations of computer security techniques (POLA, SeL4-inspired systems, hardened hardware)
  ● Automated red-teaming, vulnerability discovery
  ● Cryptographic techniques for trustworthy coordination architectures
  ● Other concrete approaches in this area

Safe Multi-Agent Scenarios

  ● Game theory addressing interactions between multiple humans, AIs, or AGIs
  ● Avoiding collusion and deception, encouraging positive-sum dynamics
  ● Principal-agent problems, Active Inference agents
  ● Other concrete approaches in this area

We encourage proposals that integrate approaches which span across multiple focus areas. 

We do not fund projects outside the four specified areas of AI safety outlined above. Common exclusions include:

  ● General AI alignment research
  ● Application-based AI technology proposals (e.g., AI for healthcare, medical research, education, or business tools)
  ● Projects in non-AI safety fields 

If you have questions please contact grants@foresight.org