Philanthropy has made systems change a fixture of its strategies, portfolios, and conferences, and yet the systems it seeks to influence keep behaving in remarkably durable ways. AI arrives at this moment as both an opportunity and a risk. Used poorly, it will harden the conventions that already work against transformative change. Used well, it can help funders see what they’ve been missing.
This paper from Jewlya Lynn and Tanya Beer names five traps that routinely constrain foundations’ systems change strategies, from designing for manageability, to failing to account for how systems defend against change, to defining success in ways that don’t last. For each trap, the authors show how AI trained in systems change frameworks can pressure-test strategies, surface hidden assumptions, and anticipate how systems defend against change. The paper also looks squarely at how AI could entrench these same traps, and at the human roles philanthropy must protect.
The window to shape how it reshapes the systems we work in is short. It can’t dismantle philanthropy’s core structures, but it can help us see them, see beyond them, and change our actions.