Kami of the Commons: Towards Designing Agentic AI to Steward the Commons

Commons suffer from neglect, free-riding, and a persistent deficit of care. Inspired by Shinto animism -- where every forest, river, and mountain has its own \emph{kami}, a spirit that inhabits and cares for that place -- we provoke: what if every commons had its own AI steward? Through a speculative design workshop where fifteen participants used Protocol Futuring, we surface both new opportunities and new dangers. Agentic AI offers the possibility of continuously supporting commons with programmable agency and care -- stewards that mediate family life as the most intimate commons, preserve collective knowledge, govern shared natural resources, and sustain community welfare. But when every commons has its own steward, second-order effects emerge: stewards contest stewards as overlapping commons collide; individuals caught between multiple stewards face new politics of care and constraint; the stewards themselves become commons requiring governance. This work opens \emph{agentive governance as commoning design material} -- a new design space for the agency, care ethics, and accountability of AI stewards of shared resources -- radically different from surveillance or optimization.