Coefficient Giving has announced a new request for proposals focused on “AI for Forecasting and Sound Reasoning,” inviting organisations and individuals to pitch innovative AI-driven projects that can enhance prediction of the future and reason more reliably.

Coefficient Giving: AI for Forecasting and Sound Reasoning

Overview

Coefficient Giving has launched a funding opportunity to support innovative AI projects that enhance human judgment, probabilistic forecasting, and reasoning. The initiative seeks to fund tools and research that improve how we predict future events and reason reliably, particularly in high-stakes or uncertain contexts.

Why It Matters

Improved Decision-Making: AI tools can structure human judgment and quantify uncertainties.
Enhanced Forecasting Accuracy: Projects that improve probabilistic predictions where data is limited are encouraged.
Sound Reasoning: AI applications can promote logical, truthful, and consistent thinking, supporting critical decisions.
Broad Applicability: While relevant to global catastrophic risk, tools developed can scale to any complex decision-making scenario.

Funding Details

Total Budget: USD 8–10 million.
Individual Awards: Typically $100,000–$1 million.
Project Duration: Six months to two years.
Eligible Applicants: Nonprofits, academic institutions, for-profits, and individual innovators worldwide.

Key Focus Areas

1. AI for Forecasting

Develop probabilistic, judgment-based forecasting models where large, clean datasets are unavailable.
Create supportive tools such as:
- Forecast explanation systems
- Decision-support models linking forecasts to actions
- Probabilistic scenario generators

2. AI for Sound Reasoning

Build AI systems aligned with logical and epistemic principles (truthfulness, consistency, meta-reasoning).
Develop practical reasoning tools such as:
- Fact-checkers
- Argument analyzers
- Arbitrators for complex decisions

3. Decision Support Applications

Focus on high-stakes, uncertain environments, including but not limited to global catastrophic risks.
Ensure improvements are generalizable to other important decision-making challenges.

Who Is Eligible?

Nonprofit organisations
Academic institutions and research labs
For-profit companies
Individual researchers and innovators
Open globally with no geographic restrictions (subject to practical and legal considerations)

How to Apply / How It Works

Concept Development: Identify a clear AI innovation for forecasting or reasoning.
Proposal Submission: Present project goals, methodology, expected outcomes, and relevance to decision-making challenges.
Budgeting: Align requested funding with project scope and duration (6 months to 2 years).
Selection Process: Proposals are reviewed based on creativity, feasibility, impact, and alignment with program goals.

Tips for Applicants

Highlight practical applications for complex or high-stakes decision-making.
Emphasize scalability of AI models and tools.
Address both forecasting and reasoning challenges explicitly.
Include clear metrics for evaluating accuracy, reliability, or reasoning quality.

FAQ

1. Who can apply?
Global applicants including nonprofits, universities, for-profits, and individuals.

2. What is the focus of funded projects?
AI tools for probabilistic forecasting, sound reasoning, and decision support.

3. What is the funding range?
USD 100,000–$1 million per project.

4. How long can projects run?
Six months to two years.

5. Are projects limited to global catastrophic risks?
No. Projects addressing general decision-making challenges are also eligible.

6. What types of tools are encouraged?
Fact-checkers, argument analyzers, forecasting support systems, scenario generators, and other AI reasoning applications.

7. Are there geographic restrictions?
No, but funding decisions consider practical and legal factors.

Conclusion

This RFP from Coefficient Giving provides a unique opportunity for innovators at the intersection of AI and decision science. By funding projects that enhance forecasting and reasoning, the initiative aims to make complex decision-making more precise, logical, and effective, with wide-ranging applications across sectors and global challenges.