Prisoner’s Dilemma

Two players independently choose to cooperate or defect. In the classic one-shot game, defection is more attractive to each individual regardless of the other’s choice, but mutual defection is worse for both than mutual cooperation.

The game-theoretic structure and repeated-game literature are highly mature. Mapping a real-world situation requires the correct payoff ordering; otherwise ‘Prisoner’s Dilemma’ is merely a metaphor.

Individually dominant defection may produce a collectively dominated equilibrium. Repetition, reputation and reciprocity change the effective strategies.

AI transfer to investigate: AI may lower the cost of contributing or monitoring, and agents can play repeated games. Whether the game remains a Prisoner’s Dilemma depends on the payoff ordering after that change.

Proposed test, not yet conducted: Create knowledge-sharing tasks with and without the payoff ordering of a Prisoner’s Dilemma. Compare cooperation at unchanged costs and after AI-driven cost reductions. Distinguish reductions that preserve the payoff ordering from those that change it, and vary expected future interaction separately.

Use game theory for model selection, not as a decorative label.

The payoff matrix is a constructed model with T > R > P > S. Defection is dominant in this one-shot game; predictions for repeated interaction or real groups require additional assumptions.

Axelrod & Hamilton (1981), The evolution of cooperation — pp. 1391–1393; Fig. 1

Axelrod and Hamilton (1981) describe the payoffs and conditions for repeated interaction. Applications to human knowledge sharing and AI interventions require separate research.