Provocations made concrete: tools, maps, and field work you can open and pull apart. Some are live, some are still deploying.
The AI stack as six layers you can pull apart, from the planet to people and institutions: what each layer is, who the players are, what goes in and comes out, and the harms, with the flows between layers drawn as pipes.
A multi-part project: an analysis of the current AI system, and a vision for AI as public utility. Part 1 and Part 2 are live now; more parts join this lineage later.
A browsable set of 298 facilitation, systems, and futures methods, the same kit used to design every Paxterra session.
A living map of the AI debate: forty-plus sources a week, each voice plotted on belief against concern, so you can watch the conversation move.
A strategic map of the responsible-AI field (regulation, frameworks, corporate practice, with a Sweden lens), paired with a risk canvas for running it with a team.
AI risks laid out by time horizon and reversibility, so a near-term nuisance and a slow irreversible one stop getting argued about in the same breath.
The full working map behind AI Futures: eighteen loops of compute, capital, talent, labour and governance, drawn with Omidyar's Systems Practice and Meadows' leverage points. Hover a node to trace the system for yourself.
A learning commons for people building on the emerging loop, connecting scattered pioneers into a community of practice rather than another course to sit through.
The work that ran in place of the AI Horizons cohort: a landscape, a risk canvas, and the sessions to sit a team down with them.
A lightweight hub connecting live funding calls to the people who need them, built for a client network as the smallest thing that tests the behaviour before anything heavier.
Three months, nineteen stakeholder interviews, a systems synthesis, and a prioritisation tool the leadership team used to pick its next bets, then kept.
Want to run one of these inside your organisation?
How we work together