Macy was explicit that holding actions are necessary and insufficient. They hold the line; they do not advance it. A movement composed only of holding actions is structurally defensive, perpetually reactive to the latest deployment, unable to build the alternative the Turning requires. Conversely, a movement that neglects holding actions while building alternatives watches the alternatives overwhelmed by the unchecked damage of the existing system.
Applied to AI, the current governance landscape is heavily weighted toward holding actions on the supply side — regulating what AI companies may build — while leaving the demand side largely unaddressed. You On AI's diagnosis of the dam deficit names this gap: almost all architecture operates on the supply side, while citizens, workers, students, and parents navigate the transition without adequate support structures.
The insufficient character of holding actions is not a critique of the work or its practitioners. Safety researchers, regulators, labor organizers, and public-interest technologists performing holding actions are doing essential work. The point is that their work becomes durable only when the other two dimensions — new structures and the shift in consciousness — are developed in parallel. Without them, holding actions are perpetual rearguard engagements that postpone defeat rather than producing victory.
Macy's framework also clarifies what makes a specific holding action effective. Effective holding actions buy time for the alternative to emerge; they do not merely slow the damage. Regulations that preserve institutional space for new organizational forms to develop are more valuable than regulations that simply block specific deployments, because they create the conditions under which the alternative can grow.
The concept emerged from Macy's environmental and peace organizing. Its three-dimensional framework was articulated most fully in Active Hope (2012), where each dimension is described in detail and their interdependence is emphasized.
Necessary and insufficient. Holding actions hold the line; they do not advance it. They require the other two dimensions to produce durable change.
Buying time, not producing alternatives. The best holding action creates institutional space for the alternative to emerge.
Supply-side and demand-side. Current AI governance is heavily weighted toward supply-side holding; demand-side structures are underdeveloped.
Reactive without the other dimensions. A movement of holding actions alone is perpetually reactive, defending ground without advancing.
Effective holding is strategic. Not every action that slows harm is equally valuable; the strategic question is whether the time bought is used to build the alternative.