February 10, 2025
Value preferences estimation and disambiguation in hybrid participatory systems
Understanding citizens’ values is essential for creating more citizen-centric policies. In participatory systems, people often make choices while also explaining the reasons behind those choices. However, choices and motivations do not always align, making it challenging to accurately identify the values that guide citizens’ decisions. In this article, Enrico Liscio explores a hybrid participatory system where AI agents interact with participants to estimate their value preferences.
Abstract
Understanding citizens’ values in participatory systems is crucial for citizen-centricpolicy-making. We envision a hybrid participatory system where participants make choicesand provide motivations for those choices, and AI agents estimate their value preferencesby interacting with them. We focus on situations where a conflict is detected betweenparticipants’ choices and motivations, and propose methods for estimating value prefer-ences while addressing detected inconsistencies by interacting with the participants. Weoperationalize the philosophical stance that “valuing is deliberatively consequential.” Thatis, if a participant’s choice is based on a deliberation of value preferences, the value prefer-ences can be observed in the motivation the participant provides for the choice. Thus, wepropose and compare value preferences estimation methods that prioritize the values esti-mated from motivations over the values estimated from choices alone. Then, we introduce adisambiguation strategy that combines Natural Language Processing and Active Learningto address the detected inconsistencies between choices and motivations. We evaluate theproposed methods on a dataset of a large-scale survey on energy transition. The resultsshow that explicitly addressing inconsistencies between choices and motivations improvesthe estimation of an individual’s value preferences. The disambiguation strategy does notshow substantial improvements when compared to similar baselines—however, we discusshow the novelty of the approach can open new research avenues and propose improvementsto address the current limitations.
To continue reading please visit:
https://doi.org/10.1613/jair.1.14958
This open access article was published in Journal of Artificial Intelligence Research 82 (2025) 819-850.
Liscio, E., Siebert, L. C., Jonker, C. M., & Murukannaiah, P. K. (2025). Value Preferences Estimation and Disambiguation in Hybrid Participatory Systems. Journal of Artificial Intelligence Research, 82, 819–850. https://doi.org/10.1613/jair.1.14958
Keywords: Natural Language Processing, Active Learning, Participatory Systems, AI Agents, Values, Policy making
Author
See also
Shared responsibility, unequal power: Mapping the limitations of multi-stakeholderism in the EU’s digital governance
By Charis Papaevangelou • Sabrina Kutscher • Natali Helberger • José van Dijck • Thomas Poell • June 01, 2026
Trading nuance for scale? Platform observability and content governance under the DSA
By Charis Papaevangelou • Fabio Votta • September 22, 2025
State, platform capitalism and infrastructural power: Microsoft's data centres in Greece 2.0
By Charis Papaevangelou • February 28, 2025
More results /
AI insiders warn it could end humanity. Does the public agree?
By Ernesto de León • Claes de Vreese • September 17, 2026
By Natali Helberger • June 16, 2026
By José van Dijck • April 30, 2026
Loss of judicial sight? The impact of assistive AI technologies on judicial perception — beyond the question of discretion
By Iris van Domselaar • Isabella Banks • September 23, 2026
By Tynke Schepers • July 09, 2026
By Corinne Cath • May 29, 2026
Fairy Tales That Make Life Harder: AI, and the social and ethical costs of believing in it
By Martijn Logtenberg • September 28, 2026
By Martijn Logtenberg • March 13, 2026
By Martijn Logtenberg • November 20, 2025
The Label Paradox: Can AI transparency create more distrust?
By Natali Helberger • September 21, 2026
By Jin Wan • Theo Araujo • Natali Helberger • Claes de Vreese • September 17, 2026
By João Pedro Quintais • September 03, 2026
