Volunteer’s Dilemma

Everyone in a group benefits if at least one person voluntarily bears a cost. If nobody does, everyone loses. Each person therefore has a reason to hope that someone else will volunteer.

The game-theoretic structure is established and has been studied experimentally. Many real-world situations in which ‘nobody does it’ are different games, however; structural fit must be tested explicitly.

Shared benefit plus a concentrated volunteering cost and the condition that one person’s action is enough creates strategic waiting. Group size and diffusion of responsibility may influence the probability of intervention.

AI transfer to investigate: AI may lower contribution costs or provide signals, but the dilemma disappears only if the effective payoffs or responsibility structure change.

Proposed test, not yet conducted: Vary group size, diffuse versus accountable assignment, and contribution cost for one necessary knowledge-maintenance action. Measure intervention probability and delay.

Not every stewardship failure is a Volunteer’s Dilemma. Use the model only when the one-volunteer threshold and payoff structure genuinely fit.

One person’s action must be enough to provide the shared benefit. Otherwise a public-goods or coordination model may fit better. A model prediction is not measured group behaviour.

Diekmann (1985), Volunteer’s Dilemma — pp. 605–610

Diekmann (1993), Cooperation in an asymmetric Volunteer’s dilemma game — abstract; pp. 75–85

Diekmann (1985, pp. 605–610) describes the model. Only the abstract of the 1993 experiment was accessible. Model predictions do not replace measurements of actual group behaviour.