This fall, Humanity AI, a collaborative of ten philanthropies including the Ford, MacArthur, Mellon, and Mozilla foundations, ran its first open call, Communities Leading on AI. The $10 million call supports U.S.-based, community-led work that gives people more power over how AI is developed, governed, and deployed, with grants from $75,000 to $1 million for projects lasting up to three years. It made a point of inviting groups that don't usually think of themselves as shaping AI, and it favors open-weight or open-source models whenever a project depends on AI models.
The call asked each proposal to pick one of three issue areas (education; labor and economy; or humanities, arts, and culture) and one of three approaches (shaping how AI is governed and used, building AI on community terms, or challenging AI that doesn't serve communities). We applied in humanities, arts, and culture, under Building AI on Community Terms. That approach gives priority to technical work on AI infrastructure, and it also leaves room for new kinds of AI tools that close a gap for a community and are designed together with that community.
The New York Botanical Garden is the lead applicant, partnering with Columbia University Libraries and with OpenRefine, which would handle project management. We're proposing an open-source, open-weight CLIP-style model that connects images and text, so a land manager or native plant gardener could photograph a plant and learn its history, its ecosystem, and the companion species and native plant communities it belongs to. A community outreach coordinator at a Hudson Valley land stewardship organization would involve those users in design decisions throughout the three-year project, and the model and methodology would be reusable by communities beyond the Hudson Valley.
If our letter of interest makes it to the next round, we will submit a full application in December. Any awarded funding would be disbursed in 2027.
Claude Opus 5.5 was used to create this summary from the letter of interest text.