Introducing CAIF's Impact Report

We're pleased to share the Cooperative AI Foundation's 2025 Impact Report, which covers activities from our inception until the end of 2024 and our initial assessments of the progress we have made. The aim with this report is to provide the reader with a clear view of what concretely CAIF is doing and why.

‍

Impact reporting is challenging for a young organization such as CAIF which is focused on interventions that typically take years to pay off in terms of measurable impact; for example, many of the research grants we have made are for projects that have not yet concluded and where it is still early to draw conclusions about their long-term impact. At this early stage, the impact report is therefore necessarily more focused on reporting activities along with our rationale for thinking that these activities will be impactful. That said, we do observe many early indications that our work seems promising.

‍

A selection of our activities from 2024, more thoroughly covered in the full report:

• We hosted our second Summer School, with 74% of attendees reporting it increased their motivation to pursue careers in cooperative AI;
• We launched our PhD fellowship program, selecting 16 fellows from 150 applications;
• We awarded new grants totaling £657,000 and revamped our grant program;
• We organized the Summer Retreat for senior researchers in Santa Cruz, where a majority of participants stated that the event seeded new collaborations and helped clarify their research priorities;
• We ran the successful Concordia Contest at NeurIPS 2024, which attracted 197 participants and 878 submissions;
• We developed an online curriculum to make cooperative AI concepts more accessible;
• We engaged productively with the multi-stakeholder consultation for the EU AI Act Code of Practice, advocating for the inclusion of multi-agent risk perspectives.

‍

While we are still in the early stages, we are encouraged by the momentum building in the cooperative AI community and the growing openness to cooperative AI ideas among policy-makers and the wider policy community. These initial learnings will inform our strategic planning and help us optimize our resource allocation and program design to maximize our contribution toward a flourishing future for all.