Cumulative Responsibility

In The You On AI Field Guide

Cumulative responsibility is not governance by committee, which Follett regarded as merely distributing the illusion of final authority among a group. Committee governance remains bounded by members' collective limitations, and the political dynamics of committee deliberation — compromises, logrolling, suppression of dissent — often produce decisions worse than those of a well-informed individual. Follett's model is the expert team, in which each member's authority derives from situated knowledge rather than hierarchical position, and the decision is produced through the integrative process.

The AI age makes cumulative responsibility more achievable because every member of the team can access AI tools that extend her situated knowledge into adjacent domains without eliminating her specific expertise. The engineer remains the engineer; AI amplifies her engineering judgment while allowing her to engage with design and strategic dimensions her bandwidth previously excluded. The designer remains the designer while engaging with technical and business dimensions. Each member's authoritative contribution becomes richer without the roles collapsing into interchangeable generalism.

The AI age also makes cumulative responsibility more urgent because the concentration of decisional authority in a single AI-augmented executive produces brittle organizations whose intelligence flows through a single bottleneck at machine speed. The alternative — distributed authority rooted in situated knowledge, integrated through collective process — produces organizations whose intelligence is genuinely collective and therefore more resilient, more accurate, and more adaptive.

The practical implementation is the structure You On AI describes as vector pods — small groups whose authority derives from situated knowledge and whose decisions emerge through integrative process. But the structure works only if the organizational culture genuinely supports distributed authority. Formal structures for cumulative responsibility fail when hierarchical override remains the practiced norm, producing consultative theater in which decisions are made above the pod and rationalized through participation rituals the pod members quickly recognize as decorative.

Origin

The concept emerged from Follett's analysis of how actual decisions get made in well-functioning organizations. She observed that behind the formal authority of senior decision-makers was always an informal network of expertise that the formal decision-maker trusted on specific questions. Cumulative responsibility made the informal network visible and legitimate, aligning formal authority with the distributed knowledge that was already shaping outcomes.

Key Ideas

Authority from situated knowledge. Each member has authority over the aspects of the situation she knows best.

Not governance by committee. The integrative process is different from committee deliberation; it is structured around reading the situation, not building consensus.

AI makes it more achievable. Tools extend each member's situated knowledge into adjacent domains without collapsing expertise into generalism.

AI makes it more urgent. Concentrated AI-augmented authority produces brittle organizations; distributed authority produces resilient ones.

Requires cultural support. Formal structures fail without genuine organizational commitment to distributed authority.