There is enormous potential to use machine learningtechniques to help philanthropic foundations evaluate the tens of thousands of proposals they receive each year. AI tools can help grantors and applicants identify promising grants and improve their resource allocations, provide suggestions and impact estimates for potential collaborations and future work, identify latent biases in existing processes, and lead to greater impact for each fund. Understanding patterns in grants data also helps grantors understand where a field’s explorations are heading, years before outcomes are published.
Philanthrobotics, an effort to advance work towards a universe of open grants data and analyses, is a collaboration between MIT Open Learning, MIT Solve, the Knowledge Futures Group, and Lever for Change around the planning, data collection, data structuring, model creation, application development, and publishing required to build quantitative tools to support philanthropy. The project, originated in 2020 with support from Lever for Change, has begun to:
· collect and structure grant application & decision data;
· link collected data to information sources & visualizations;
· analyze statistics from grant applications & grant reports;
· run machine-learning analyses of the collected data; and
· help the broader community convene and publish their work
How can we improve the commons of open grants data? How can this improve grantmaking for grantors and recipients? How can we better add data and context to proposals we receive, and make that accessible? And how can we make machine evaluation useful for the field? What limits to these approaches need consideration – and how can AI help us identify existing biases, rather than reifying them?
Join us to share your approaches to this, to learn from other groups with similar initiatives, and to plan next steps.
