Story perspectives

Memory Cues Enable Cooperation in Prisoner’s Dilemma Simulations

5/19/2026

Summary

Story summary

Alexandre V. Morozov, Rutgers physics professor, and Alexander Feigel, Hebrew University researcher, published a PNAS paper on 22-May-2026 showing memory-based opponent recognition can evolve cooperation in repeated Prisoner’s Dilemma games.
Their model adds a memory system that shifts payoffs, letting cooperative strategies outcompete cheating.
Neural-network simulations showed agents tracking opponents achieve cooperation without kin selection, suggesting microbes or insects could cooperate via opponent-specific cues and challenging view that cheating always wins.

1 of 1

Story summary

Alexandre V. Morozov, Rutgers physics professor, and Alexander Feigel, Hebrew University researcher, published a PNAS paper on 22-May-2026 showing memory-based opponent recognition can evolve cooperation in repeated Prisoner’s Dilemma games.
Their model adds a memory system that shifts payoffs, letting cooperative strategies outcompete cheating.
Neural-network simulations showed agents tracking opponents achieve cooperation without kin selection, suggesting microbes or insects could cooperate via opponent-specific cues and challenging view that cheating always wins.