Generative AI for climate governance and acceptability-constrained policy design

Ajaykumar Manivannan; Viktoria Spaiser; Tristan Cann; James Evans; Jordan Everall; Max Falkenberg; David Garcia; Weisi Guo; Rico Herzog; Ilona Otto; Yannick Oswald; Nicolò Pagan; Max Pellert; Charlie Pilgrim; Carlos Rodriguez-Pardo; Indira Sen; Alexander Vezhnevets

doi:10.1038/s44168-026-00362-6

Generative AI for climate governance and acceptability-constrained policy design

Ajaykumar Manivannan^*, Viktoria Spaiser^*, Tristan Cann, James Evans, Jordan Everall, Max Falkenberg, David Garcia, Weisi Guo, Rico Herzog, Ilona Otto, Yannick Oswald, Nicolò Pagan, Max Pellert, Charlie Pilgrim, Carlos Rodriguez-Pardo, Indira Sen, Alexander Vezhnevets

^*Corresponding author for this work

Digital Urban Cultures

Abstract

Climate policies often fail when they clash with cultural values, social identities, and fairness perceptions. We propose Acceptability-Constrained Climate Policy Design (ACCPD), using large language models as "cultural world models" to simulate public responses before implementation. By embedding LLMs in generative agent-based models and physical system simulators, ACCPD aims to enable policymakers to co-optimize for climate-policy efficacy and social legitimacy. We discuss methodological limitations regarding representation and LLM opacity.

Original language	English
Article number	37
Number of pages	8
Journal	npj Climate Action
Volume	5
Issue number	1
DOIs	https://doi.org/10.1038/s44168-026-00362-6
Publication status	Published - 2026

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 13 Climate Action

Free Keywords

Generative grammar
Representation (politics)
Corporate governance
Public policy
Climate policy
Climate governance
Climate change

Access to Document

10.1038/s44168-026-00362-6Licence: CC BY

Cite this

Manivannan, A., Spaiser, V., Cann, T., Evans, J., Everall, J., Falkenberg, M., Garcia, D., Guo, W., Herzog, R., Otto, I., Oswald, Y., Pagan, N., Pellert, M., Pilgrim, C., Rodriguez-Pardo, C., Sen, I., & Vezhnevets, A. (2026). Generative AI for climate governance and acceptability-constrained policy design. npj Climate Action, 5(1), Article 37. https://doi.org/10.1038/s44168-026-00362-6

@article{fb7c448f92b048e5ba706e9343e7f5fd,

title = "Generative AI for climate governance and acceptability-constrained policy design",

abstract = "Climate policies often fail when they clash with cultural values, social identities, and fairness perceptions. We propose Acceptability-Constrained Climate Policy Design (ACCPD), using large language models as {"}cultural world models{"} to simulate public responses before implementation. By embedding LLMs in generative agent-based models and physical system simulators, ACCPD aims to enable policymakers to co-optimize for climate-policy efficacy and social legitimacy. We discuss methodological limitations regarding representation and LLM opacity.",

keywords = "Generative grammar, Representation (politics), Corporate governance, Public policy, Climate policy, Climate governance, Climate change",

author = "Ajaykumar Manivannan and Viktoria Spaiser and Tristan Cann and James Evans and Jordan Everall and Max Falkenberg and David Garcia and Weisi Guo and Rico Herzog and Ilona Otto and Yannick Oswald and Nicol{\`o} Pagan and Max Pellert and Charlie Pilgrim and Carlos Rodriguez-Pardo and Indira Sen and Alexander Vezhnevets",

year = "2026",

doi = "10.1038/s44168-026-00362-6",

language = "English",

volume = "5",

journal = "npj Climate Action",

issn = "2731-9814",

number = "1",

}

Manivannan, A, Spaiser, V, Cann, T, Evans, J, Everall, J, Falkenberg, M, Garcia, D, Guo, W, Herzog, R, Otto, I, Oswald, Y, Pagan, N, Pellert, M, Pilgrim, C, Rodriguez-Pardo, C, Sen, I & Vezhnevets, A 2026, 'Generative AI for climate governance and acceptability-constrained policy design', npj Climate Action, vol. 5, no. 1, 37. https://doi.org/10.1038/s44168-026-00362-6

TY - JOUR

T1 - Generative AI for climate governance and acceptability-constrained policy design

AU - Manivannan, Ajaykumar

AU - Spaiser, Viktoria

AU - Cann, Tristan

AU - Evans, James

AU - Everall, Jordan

AU - Falkenberg, Max

AU - Garcia, David

AU - Guo, Weisi

AU - Herzog, Rico

AU - Otto, Ilona

AU - Oswald, Yannick

AU - Pagan, Nicolò

AU - Pellert, Max

AU - Pilgrim, Charlie

AU - Rodriguez-Pardo, Carlos

AU - Sen, Indira

AU - Vezhnevets, Alexander

PY - 2026

Y1 - 2026

N2 - Climate policies often fail when they clash with cultural values, social identities, and fairness perceptions. We propose Acceptability-Constrained Climate Policy Design (ACCPD), using large language models as "cultural world models" to simulate public responses before implementation. By embedding LLMs in generative agent-based models and physical system simulators, ACCPD aims to enable policymakers to co-optimize for climate-policy efficacy and social legitimacy. We discuss methodological limitations regarding representation and LLM opacity.

AB - Climate policies often fail when they clash with cultural values, social identities, and fairness perceptions. We propose Acceptability-Constrained Climate Policy Design (ACCPD), using large language models as "cultural world models" to simulate public responses before implementation. By embedding LLMs in generative agent-based models and physical system simulators, ACCPD aims to enable policymakers to co-optimize for climate-policy efficacy and social legitimacy. We discuss methodological limitations regarding representation and LLM opacity.

KW - Generative grammar

KW - Representation (politics)

KW - Corporate governance

KW - Public policy

KW - Climate policy

KW - Climate governance

KW - Climate change

U2 - 10.1038/s44168-026-00362-6

DO - 10.1038/s44168-026-00362-6

M3 - Journal Article

SN - 2731-9814

VL - 5

JO - npj Climate Action

JF - npj Climate Action

IS - 1

M1 - 37

ER -