Cooperative AI Foundation Launches 2025 Research Grants With New Priorities and Early-Career Track

The Cooperative AI Founda

The Cooperative AI Foundation (CAIF) has officially launched its 2025 Research Grant Program, signaling a major push to advance the field of cooperative artificial intelligence—with sharpened priorities, a more structured application process, and a brand-new early-career track designed to nurture the next generation of AI researchers.

While the first round of applications has already closed, the initiative has drawn significant attention from the global research community, given CAIF’s increasingly focused mission: to ensure that AI systems can cooperate safely and effectively with humans and with each other.

What’s New in 2025?

CAIF’s 2025 program introduces several updates aimed at improving both impact and accessibility:

• Refined Priority Areas: High priority is placed on:

  1. Understanding and Evaluating Cooperation-Relevant Capabilities

  2. Understanding and Evaluating Cooperation-Relevant Propensities

  These topics explore how to measure and predict whether AI systems will act cooperatively in multi-agent or human-AI environments.

• Clearer Scope and Guidelines:
  The foundation has published detailed technical and methodological guidelines, benchmark expectations, and proposal writing standards to help applicants better align with funding goals. Only research proposals related to the listed themes will be considered.

• New Early-Career Track:
  Designed for researchers within ~2–3 years of completing a PhD (or an equivalent stage), this track provides up to £100,000 for 12-month individual projects. Importantly, these proposals are evaluated not just on research impact, but also on how the grant could accelerate the applicant’s career trajectory.

Key Research Areas for 2025

Applicants must align their proposals with at least one of CAIF’s eight core research areas:

1. Cooperation-Relevant Capabilities (High Priority)

2. Cooperation-Relevant Propensities (High Priority)

3. Incentivizing Cooperation Among AI Agents

4. AI for Facilitating Human Cooperation

5. Collusion

6. Monitoring and Controlling Dynamic Agent Networks

7. Information Asymmetries and Transparency

8. Multi-Agent Security

Projects addressing systemic risks, emergent behavior, and real-world coordination failures are strongly encouraged. 

Application & Selection Process

CAIF follows a two-stage application process:

1. Pre-Proposal:
   A short application (~2 pages) including project summary, applicant bio, and budget estimate. No feedback is provided for rejected pre-proposals.

2. Full Proposal (Invite-Only):
   Selected applicants receive detailed feedback and can revise their submission before final decisions are made (expected by May 2025).

While the foundation received a high volume of strong submissions in its first 2025 call, it has not yet confirmed a second round due to funding limitations. A “Next Deadline” is to be confirmed (TBC).

Funding Details

• Project Duration: Up to 2 years

• Minimum Budget: £10,000

• No Fixed Max, but cost-effectiveness is a major criterion

• Eligible Costs: Research staff, compute, travel, publications

• Indirect Costs: Capped at 10%

• Teaching salaries are not covered

Importantly, applicants may be unaffiliated and based anywhere globally—though proposals from regions with low CPI may undergo extended due diligence.

Why It Matters

As AI systems grow increasingly autonomous and embedded in real-world decision-making, ensuring they cooperate ethically, safely, and efficiently becomes a foundational challenge. CAIF’s 2025 program is a concrete step toward building the tools, models, and understanding needed to address this challenge head-on.

The foundation also continues to build community resources, such as the Introduction to Cooperative AI Curriculum (currently in development), and invites those with educational or advocacy project ideas to reach out independently.

Learn More & Stay Updated

• View the 2025 Research Call: https://www.cooperativeai.com/grants/2025

• Explore previous funded projects: cooperativeai.com

• Subscribe for updates or follow on Twitter/X: @cooperativeAI

Final Note:
Whether you’re an academic researcher, an AI safety enthusiast, or an early-career innovator, CAIF’s evolving grant program represents an exciting opportunity to shape how intelligent systems cooperate, not compete, in the age of artificial intelligence.