Deep reinforcement learning can promote sustainable human behaviour in a common-pool resource problem

Abstract A canonical social dilemma arises when resources are allocated to people, who can either reciprocate with interest or keep the proceeds. The right resource allocation mechanisms can encourage levels of reciprocation that sustain the commons. Here, in an iterated multiplayer trust game, we u...

Full description

Saved in:
Bibliographic Details
Main Authors: Raphael Koster, Miruna Pîslar, Andrea Tacchetti, Jan Balaguer, Leqi Liu, Romuald Elie, Oliver P. Hauser, Karl Tuyls, Matt Botvinick, Christopher Summerfield
Format: Article
Language:English
Published: Nature Portfolio 2025-03-01
Series:Nature Communications
Online Access:https://doi.org/10.1038/s41467-025-58043-7
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849389942966321152
author Raphael Koster
Miruna Pîslar
Andrea Tacchetti
Jan Balaguer
Leqi Liu
Romuald Elie
Oliver P. Hauser
Karl Tuyls
Matt Botvinick
Christopher Summerfield
author_facet Raphael Koster
Miruna Pîslar
Andrea Tacchetti
Jan Balaguer
Leqi Liu
Romuald Elie
Oliver P. Hauser
Karl Tuyls
Matt Botvinick
Christopher Summerfield
author_sort Raphael Koster
collection DOAJ
description Abstract A canonical social dilemma arises when resources are allocated to people, who can either reciprocate with interest or keep the proceeds. The right resource allocation mechanisms can encourage levels of reciprocation that sustain the commons. Here, in an iterated multiplayer trust game, we use deep reinforcement learning (RL) to design a social planner that promotes sustainable contributions from human participants. We first trained neural networks to behave like human players, creating a stimulated economy that allows us to study the dynamics of receipt and reciprocation. We use RL to train a mechanism to maximise aggregate return to players. The RL mechanism discovers a redistributive policy that leads to a large but also more equal surplus. The mechanism outperforms baseline mechanisms by conditioning its generosity on available resources and temporarily sanctioning defectors. Examining the RL policy allows us to develop a similar but explainable mechanism that is more popular among players.
format Article
id doaj-art-d1b0a6dc406a496cb550eccf93247fc5
institution Kabale University
issn 2041-1723
language English
publishDate 2025-03-01
publisher Nature Portfolio
record_format Article
series Nature Communications
spelling doaj-art-d1b0a6dc406a496cb550eccf93247fc52025-08-20T03:41:49ZengNature PortfolioNature Communications2041-17232025-03-0116111310.1038/s41467-025-58043-7Deep reinforcement learning can promote sustainable human behaviour in a common-pool resource problemRaphael Koster0Miruna Pîslar1Andrea Tacchetti2Jan Balaguer3Leqi Liu4Romuald Elie5Oliver P. Hauser6Karl Tuyls7Matt Botvinick8Christopher Summerfield9Google DeepMindGoogle DeepMindGoogle DeepMindGoogle DeepMindGoogle DeepMindGoogle DeepMindUniversity of ExeterGoogle DeepMindGoogle DeepMindUniversity of OxfordAbstract A canonical social dilemma arises when resources are allocated to people, who can either reciprocate with interest or keep the proceeds. The right resource allocation mechanisms can encourage levels of reciprocation that sustain the commons. Here, in an iterated multiplayer trust game, we use deep reinforcement learning (RL) to design a social planner that promotes sustainable contributions from human participants. We first trained neural networks to behave like human players, creating a stimulated economy that allows us to study the dynamics of receipt and reciprocation. We use RL to train a mechanism to maximise aggregate return to players. The RL mechanism discovers a redistributive policy that leads to a large but also more equal surplus. The mechanism outperforms baseline mechanisms by conditioning its generosity on available resources and temporarily sanctioning defectors. Examining the RL policy allows us to develop a similar but explainable mechanism that is more popular among players.https://doi.org/10.1038/s41467-025-58043-7
spellingShingle Raphael Koster
Miruna Pîslar
Andrea Tacchetti
Jan Balaguer
Leqi Liu
Romuald Elie
Oliver P. Hauser
Karl Tuyls
Matt Botvinick
Christopher Summerfield
Deep reinforcement learning can promote sustainable human behaviour in a common-pool resource problem
Nature Communications
title Deep reinforcement learning can promote sustainable human behaviour in a common-pool resource problem
title_full Deep reinforcement learning can promote sustainable human behaviour in a common-pool resource problem
title_fullStr Deep reinforcement learning can promote sustainable human behaviour in a common-pool resource problem
title_full_unstemmed Deep reinforcement learning can promote sustainable human behaviour in a common-pool resource problem
title_short Deep reinforcement learning can promote sustainable human behaviour in a common-pool resource problem
title_sort deep reinforcement learning can promote sustainable human behaviour in a common pool resource problem
url https://doi.org/10.1038/s41467-025-58043-7
work_keys_str_mv AT raphaelkoster deepreinforcementlearningcanpromotesustainablehumanbehaviourinacommonpoolresourceproblem
AT mirunapislar deepreinforcementlearningcanpromotesustainablehumanbehaviourinacommonpoolresourceproblem
AT andreatacchetti deepreinforcementlearningcanpromotesustainablehumanbehaviourinacommonpoolresourceproblem
AT janbalaguer deepreinforcementlearningcanpromotesustainablehumanbehaviourinacommonpoolresourceproblem
AT leqiliu deepreinforcementlearningcanpromotesustainablehumanbehaviourinacommonpoolresourceproblem
AT romualdelie deepreinforcementlearningcanpromotesustainablehumanbehaviourinacommonpoolresourceproblem
AT oliverphauser deepreinforcementlearningcanpromotesustainablehumanbehaviourinacommonpoolresourceproblem
AT karltuyls deepreinforcementlearningcanpromotesustainablehumanbehaviourinacommonpoolresourceproblem
AT mattbotvinick deepreinforcementlearningcanpromotesustainablehumanbehaviourinacommonpoolresourceproblem
AT christophersummerfield deepreinforcementlearningcanpromotesustainablehumanbehaviourinacommonpoolresourceproblem