Human-centred mechanism design with democratic AI
dc.contributor.author | Koster, R | |
dc.contributor.author | Balaguer, J | |
dc.contributor.author | Tacchetti, A | |
dc.contributor.author | Weinstein, A | |
dc.contributor.author | Zhu, T | |
dc.contributor.author | Hauser, OP | |
dc.contributor.author | Williams, D | |
dc.contributor.author | Campbell-Gillingham, L | |
dc.contributor.author | Thacker, P | |
dc.contributor.author | Botvinick, M | |
dc.contributor.author | Summerfield, C | |
dc.date.accessioned | 2022-05-11T08:18:02Z | |
dc.date.issued | 2022-07-04 | |
dc.date.updated | 2022-05-10T20:46:23Z | |
dc.description.abstract | Building artificial intelligence (AI) that aligns with human values is an unsolved problem. Here, we developed a human-in-the-loop research pipeline called Democratic AI, in which reinforcement learning is used to design a social mechanism that humans prefer by majority. A large group of humans played an online investment game that involved deciding whether to keep a monetary endowment or to share it with others for collective benefit. Shared revenue was returned to players under two different redistribution mechanisms, one designed by the AI and the other by humans. The AI discovered a mechanism that redressed initial wealth imbalance, sanctioned free riders, and successfully won the majority vote. By optimizing for human preferences, Democratic AI offers a proof of concept for value-aligned policy innovation | en_GB |
dc.identifier.citation | Published online 4 July 2022 | en_GB |
dc.identifier.doi | 10.1038/s41562-022-01383-x | |
dc.identifier.uri | http://hdl.handle.net/10871/129585 | |
dc.identifier | ORCID: 0000-0002-9282-0801 (Hauser, Oliver P) | |
dc.language.iso | en | en_GB |
dc.publisher | Nature Research | en_GB |
dc.relation.url | https://github.com/deepmind/hcmd_dai | en_GB |
dc.rights | © The Author(s) 2022. Open Access. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. | |
dc.title | Human-centred mechanism design with democratic AI | en_GB |
dc.type | Article | en_GB |
dc.date.available | 2022-05-11T08:18:02Z | |
dc.identifier.issn | 2397-3374 | |
dc.description | This is the final version. Available on open access from Nature Research via the DOI in this record | en_GB |
dc.description | Data availability: All human data is available at https://github.com/deepmind/hcmd_dai | en_GB |
dc.description | Code availability: Code for reproducing figures is available at https://github.com/deepmind/hcmd_dai | en_GB |
dc.identifier.journal | Nature Human Behaviour | en_GB |
dc.rights.uri | http://creativecommons.org/licenses/by/4.0/ | en_GB |
dcterms.dateAccepted | 2022-05-10 | |
rioxxterms.version | VoR | en_GB |
rioxxterms.licenseref.startdate | 2022-05-10 | |
rioxxterms.type | Journal Article/Review | en_GB |
refterms.dateFCD | 2022-05-10T20:46:28Z | |
refterms.versionFCD | AM | |
refterms.dateFOA | 2022-07-06T12:05:19Z | |
refterms.panel | C | en_GB |
Files in this item
This item appears in the following Collection(s)
Except where otherwise noted, this item's licence is described as © The Author(s) 2022. Open Access. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.