pith. sign in

arxiv: 2605.16291 · v1 · pith:APDC7MRYnew · submitted 2026-04-14 · 💻 cs.CY · cs.AI· cs.GT

AI of the People, by the People, for the People: A Social Choice Approach to Collective Control of Artificial Intelligence

Paul Anton Bachmann , Niclas Boehmer , Lukas Daniel Klausner , Martin Lackner This is my paper

Pith reviewed 2026-05-21 01:36 UTC · model grok-4.3

classification 💻 cs.CY cs.AIcs.GT
keywords social choice theorycollective controlAI governancemachine learning pipelinepreference aggregationAI alignmentdemocratic control
0
0 comments X

The pith

Social choice theory can guide collective societal control over AI systems across the entire development process.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces collective control of artificial intelligence as a way to involve groups of people in shaping AI through social choice methods. It claims this input should happen at multiple stages including selecting training data, defining objectives, and ensuring alignment with human values. This approach matters because it moves beyond high-level policy to practical mechanisms that use mathematical criteria for fairness in decision-making. Readers interested in democratic technology would see value in treating AI design as a series of collective choice problems solvable with established voting and aggregation tools.

Core claim

We propose a new approach grounded in social choice theory, which we term collective control of artificial intelligence. We argue that collective input can and should be incorporated at multiple points across the ML development pipeline, from data collection through objective design to alignment. We further demonstrate that social choice provides a well-suited modelling language for the treatment of collective input across all stages and that its axiomatic methodology yields principled criteria for evaluating various control mechanisms. Overall, our conceptual contribution provides a mathematically grounded framework to implement and analyse collective control of AI systems.

What carries the argument

collective control of artificial intelligence, which applies social choice theory to aggregate public input for decisions at each stage of the machine learning pipeline

If this is right

  • Control mechanisms for data collection can be evaluated for properties like representativeness using social choice axioms.
  • Objective design processes can incorporate diverse societal preferences through preference aggregation methods.
  • Alignment techniques become subject to formal criteria for collective approval rather than solely expert judgment.
  • The framework allows analysis of trade-offs between different points of intervention in AI development.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Designers could prototype voting-based systems for choosing AI training datasets in public experiments.
  • This view links AI ethics to classic problems of fair division and resource allocation studied in economics.
  • Future work might test whether axiomatic guarantees translate to better real-world acceptance of AI systems.

Load-bearing premise

Social choice theory provides a well-suited modelling language for the treatment of collective input across all stages of the ML pipeline and its axiomatic methodology yields principled criteria for evaluating control mechanisms.

What would settle it

An empirical demonstration that a control mechanism violating core social choice axioms produces better societal acceptance or technical performance than one satisfying those axioms.

Figures

Figures reproduced from arXiv: 2605.16291 by Lukas Daniel Klausner, Martin Lackner, Niclas Boehmer, Paul Anton Bachmann.

Figure 1
Figure 1. Figure 1: A sketched ML development process; collective input can be meaningfully incorporated at each [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: A sketch of our mathematical model, where we view an AI system [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
read the original abstract

With the growing adoption of AI systems, reasoning about how society can exert control over AI becomes an increasingly urgent problem. Existing work on democratic control largely focuses on macro-level governance. In contrast, we propose a new approach grounded in social choice theory, which we term collective control of artificial intelligence. We argue that collective input can and should be incorporated at multiple points across the ML development pipeline, from data collection through objective design to alignment. We further demonstrate that social choice provides a well-suited modelling language for the treatment of collective input across all stages and that its axiomatic methodology yields principled criteria for evaluating various control mechanisms. Overall, our conceptual contribution provides a mathematically grounded framework to implement and analyse collective control of AI systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper proposes a new approach called 'collective control of artificial intelligence' grounded in social choice theory to address societal control over AI systems. It contrasts this with existing macro-level governance work and argues that collective input should be incorporated at multiple stages of the ML development pipeline, including data collection, objective design, and alignment. The authors claim that social choice theory supplies a well-suited modeling language for collective input across these stages and that its axiomatic methodology yields principled criteria for evaluating control mechanisms, providing an overall mathematically grounded framework.

Significance. If the central claim can be substantiated by mapping social choice axioms to concrete, non-vacuous properties of control mechanisms at specific ML stages, the work would offer a valuable bridge between social choice theory and AI governance. This could enable more rigorous, principle-based design and analysis of collective decision processes in AI development, extending beyond high-level policy discussions. The conceptual framing is timely given growing AI adoption, though its immediate significance is constrained by the absence of explicit derivations or examples in the current manuscript.

major comments (2)
  1. [Abstract] Abstract: The assertion that social choice theory 'yields principled criteria for evaluating various control mechanisms' is load-bearing for the central contribution but is not supported by any concrete mapping. No example is given of how a specific axiom (e.g., Pareto efficiency, independence of irrelevant alternatives, or strategy-proofness) would rank or rule out a control mechanism at a named pipeline stage such as objective design or alignment, leaving the criteria at a general level without engaging ML-specific features like high-dimensional objectives or noisy feedback.
  2. [Abstract] Abstract: The claim that social choice provides a 'well-suited modelling language for the treatment of collective input across all stages' is asserted without addressing potential mismatches, such as the fact that many ML stages rely on non-preference-based signals (e.g., gradient updates from human feedback) rather than explicit preference aggregation; a demonstration that standard axioms extend non-trivially to these settings is needed to substantiate suitability.
minor comments (2)
  1. The introduction of the term 'collective control of artificial intelligence' as a new entity would benefit from an explicit comparison table or paragraph distinguishing it from related concepts such as participatory AI, democratic AI, or value alignment.
  2. Consider adding a diagram in the main text that maps social choice mechanisms to specific pipeline stages to improve clarity of the proposed framework.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments, which identify key areas where the manuscript's central claims would benefit from greater concreteness. We address each major comment in turn and indicate the revisions planned for the next version of the manuscript.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The assertion that social choice theory 'yields principled criteria for evaluating various control mechanisms' is load-bearing for the central contribution but is not supported by any concrete mapping. No example is given of how a specific axiom (e.g., Pareto efficiency, independence of irrelevant alternatives, or strategy-proofness) would rank or rule out a control mechanism at a named pipeline stage such as objective design or alignment, leaving the criteria at a general level without engaging ML-specific features like high-dimensional objectives or noisy feedback.

    Authors: We agree that the abstract states the claim at a general level and that a concrete mapping would strengthen the presentation. The manuscript develops the overall framework and indicates how axioms can inform mechanism evaluation at different pipeline stages, but it does not supply the detailed, stage-specific illustrations requested. We will therefore revise by inserting a new illustrative subsection that maps Pareto efficiency to objective design, showing how the axiom rules out aggregation procedures that discard unanimous preferences over particular features and discussing its interaction with high-dimensional objective spaces. revision: yes

  2. Referee: [Abstract] Abstract: The claim that social choice provides a 'well-suited modelling language for the treatment of collective input across all stages' is asserted without addressing potential mismatches, such as the fact that many ML stages rely on non-preference-based signals (e.g., gradient updates from human feedback) rather than explicit preference aggregation; a demonstration that standard axioms extend non-trivially to these settings is needed to substantiate suitability.

    Authors: The referee correctly notes that the suitability claim would be more robust if potential mismatches with non-explicit signals were addressed explicitly. The manuscript treats collective input broadly, including implicit signals derived from feedback, but does not demonstrate non-trivial axiom extensions in detail. In revision we will add a short discussion showing how strategy-proofness can be adapted to noisy human-feedback settings in the alignment stage, thereby supplying a concrete criterion for robust aggregation under gradient-based updates. revision: yes

Circularity Check

0 steps flagged

No significant circularity; proposal applies established external theory

full rationale

The paper's central contribution is a conceptual framework that applies social choice theory—an independently developed field with its own axiomatic literature—to stages of the ML pipeline. The abstract and provided text present this as an argument for suitability and for the value of axiomatic criteria, without any equations, fitted parameters, self-citations that bear the load of a uniqueness claim, or reductions that equate a derived result to its own inputs by construction. No self-definitional loops, renamed empirical patterns, or smuggled ansatzes appear in the given material. The derivation chain therefore remains self-contained against external benchmarks from social choice theory.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

Abstract-only review limits visibility into specific parameters or entities; the proposal appears to rest on standard social choice axioms being transferable to AI pipeline decisions.

axioms (1)
  • domain assumption Social choice theory axioms can be applied to model collective preferences over AI development choices.
    Invoked when claiming social choice is well-suited for collective input at multiple pipeline stages.
invented entities (1)
  • Collective control of artificial intelligence no independent evidence
    purpose: Framework for incorporating societal input into AI via social choice methods.
    New term introduced to organize the proposal; no independent evidence or falsifiable prediction provided in abstract.

pith-pipeline@v0.9.0 · 5667 in / 1160 out tokens · 50524 ms · 2026-05-21T01:36:21.752999+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

106 extracted references · 106 canonical work pages · 2 internal anchors

  1. [1]

    Dynamic Fairness-Aware Recommendation Through Multi-Agent Social Choice.ACM Trans

    Amanda Aird, Paresha Farastu, Joshua Sun, Elena Stefancov ´a, Cassidy All, Amy V oida, Nicholas Mattei, and Robin Burke. Dynamic Fairness-Aware Recommendation Through Multi-Agent Social Choice.ACM Trans. Recomm. Syst., 3(2), 2025. doi: 10.1145/3690653

    work page doi:10.1145/3690653 2025

  2. [2]

    Multi-Winner Approval V oting Goes Epistemic

    Tahar Allouche, J ´erˆome Lang, and Florian Yger. Multi-Winner Approval V oting Goes Epistemic. In Proceedings of the 38th Conference on Uncertainty in Artificial Intelligence, UAI 2022/Proceedings of Machine Learning Research 180, pages 75–84, Cambridge, MA, 2022. JMLR and Microtome Publishing. URLhttps://proceedings.mlr.press/v180/allouche22a.html

    work page 2022

  3. [3]

    Parashar et al

    Saleema Amershi, Andrew Begel, Christian Bird, Robert DeLine, Harald Gall, Ece Kamar, Nachi- appan Nagappan, Besmira Nushi, and Thomas Zimmermann. Software Engineering for Machine 16 Learning: A Case Study. In2019 IEEE/ACM 41st International Conference on Software Engineer- ing: Software Engineering in Practice, ICSE-SEIP 2019, pages 291–300, Washington, ...

    work page doi:10.1109/icse-seip.2019.00042 2019

  4. [4]

    Machine Ethics: Creating an Ethical Intelligent Agent

    Michael Anderson and Susan Leigh Anderson. Machine Ethics: Creating an Ethical Intelligent Agent. AI Mag., 28(4):15–26, 2007. doi: 10.1609/AIMAG.v28i4.2065

    work page doi:10.1609/aimag.v28i4.2065 2007

  5. [5]

    The Reanimation of Pseudoscience in Machine Learning and Its Ethical Repercussions.Patterns, 5(9), 2024

    Mel Andrews, Andrew Smart, and Abeba Birhane. The Reanimation of Pseudoscience in Machine Learning and Its Ethical Repercussions.Patterns, 5(9), 2024. doi: 10.1016/J.PATTER.2024.101027

    work page doi:10.1016/j.patter.2024.101027 2024

  6. [6]

    Learning to Elect

    Cem Anil and Xuchan Bao. Learning to Elect. InAdvances in Neural Information Pro- cessing Systems 34, NeurIPS 2021, pages 8006–8017, Red Hook, NY , 2021. Curran As- sociates. URLhttps://proceedings.neurips.cc/paper_files/paper/2021/ hash/42d6c7d61481d1c21bd1635f59edae05-Abstract.html

    work page 2021

  7. [7]

    AI Magazine , month = mar, pages =

    Lora Aroyo and Chris Welty. Truth is a Lie: Crowd Truth and the Seven Myths of Human Annotation. AI Mag., 36(1):15–24, 2015. doi: 10.1609/AIMAG.v36i1.2564

    work page doi:10.1609/aimag.v36i1.2564 2015

  8. [8]

    Crowd- sourcing for Multiple-Choice Question Answering.Proc

    Bahadir Ismail Aydin, Yavuz Selim Yilmaz, Yaliang Li, Qi Li, Jing Gao, and Murat Demirbas. Crowd- sourcing for Multiple-Choice Question Answering.Proc. AAAI Conf. Artif. Intell., 28(2):2946–2953,

    work page

  9. [9]

    doi: 10.1609/AAAI.v28i2.19016

    work page doi:10.1609/aaai.v28i2.19016

  10. [10]

    Big Tech, Algorithmic Power, and Democratic Control.J

    U ˘gur Aytac ¸. Big Tech, Algorithmic Power, and Democratic Control.J. Politics, 86(4):1431–1445,

    work page

  11. [11]

    doi: 10.1007/S43681-025-00734-4

    Nasser Bahrami.AIgemony: Power Dynamics, Dominant Narratives, and Colonisation.AI Ethics, 5 (5):5081–5103, 2025. doi: 10.1007/S43681-025-00734-4

    work page doi:10.1007/s43681-025-00734-4 2025

  12. [12]

    Bowman, Zac Hatfield-Dodds, Ben Mann, Dario Amodei, Nicholas Joseph, Sam McCandlish, Tom Brown, and Jared Kaplan

    Yuntao Bai, Saurav Kadavath, Sandipan Kundu, Amanda Askell, Jackson Kernion, Andy Jones, Anna Chen, Anna Goldie, Azalia Mirhoseini, Cameron McKinnon, Carol Chen, Catherine Olsson, Christo- pher Olah, Danny Hernandez, Dawn Drain, Deep Ganguli, Dustin Li, Eli Tran-Johnson, Ethan Perez, Jamie Kerr, Jared Mueller, Jeffrey Ladish, Joshua Landau, Kamal Ndousse,...

    work page 2022

  13. [13]

    How to Regulate Large Language Models for Responsible AI.IEEE Trans

    Jose Berengueres. How to Regulate Large Language Models for Responsible AI.IEEE Trans. Tech- nol. Soc., 5(2):191–197, 2024. doi: 10.1109/TTS.2024.3403681

    work page doi:10.1109/tts.2024.3403681 2024

  14. [14]

    Beyond Individual Accountability: (Re-)Asserting Democratic Control of AI

    Daniel James Bogiatzis-Gibbons. Beyond Individual Accountability: (Re-)Asserting Democratic Control of AI. InProceedings of the 2024 ACM Conference on Fairness, Accountability, and Trans- parency, FAccT ’24, pages 74–84, New York, NY , 2024. ACM. doi: 10.1145/3630106.3658541

    work page doi:10.1145/3630106.3658541 2024

  15. [15]

    Collective Choice under Dichotomous Pref- erences.J

    Anna Bogomolnaia, Herv ´e Moulin, and Richard Stong. Collective Choice under Dichotomous Pref- erences.J. Econ. Theory, 122(2):165–184, 2005. doi: 10.1016/J.JET.2004.05.005

    work page doi:10.1016/j.jet.2004.05.005 2005

  16. [16]

    Cambridge University Press, Cambridge, 2016

    Felix Brandt, Vincent Conitzer, Ulle Endriss, J ´erˆome Lang, and Ariel Procaccia, editors.Handbook of Computational Social Choice. Cambridge University Press, Cambridge, 2016. ISBN 978-1-107- 06043-2. 17

    work page 2016

  17. [17]

    Aggregating Complex Annotations via Merging and Match- ing

    Alexander Braylan and Matthew Lease. Aggregating Complex Annotations via Merging and Match- ing. InProceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Min- ing, KDD ’21, pages 86–94, New York, NY , 2021. ACM. doi: 10.1145/3447548.3467411

    work page doi:10.1145/3447548.3467411 2021

  18. [18]

    Multiwin- ner Elections with Diversity Constraints.Proc

    Robert Bredereck, Piotr Faliszewski, Ayumi Igarashi, Martin Lackner, and Piotr Skowron. Multiwin- ner Elections with Diversity Constraints.Proc. AAAI Conf. Artif. Intell., 32(1):933–940, 2024. doi: 10.1609/AAAI.v32i1.11457

    work page doi:10.1609/aaai.v32i1.11457 2024

  19. [19]

    V oting: A Machine Learning Approach.Eur

    D ´avid Burka, Clemens Puppe, L ´aszl´o Szepesv´ary, and Attila Tasn´adi. V oting: A Machine Learning Approach.Eur. J. Oper. Res., 299(3):1003–1017, 2022. doi: 10.1016/J.EJOR.2021.10.005

    work page doi:10.1016/j.ejor.2021.10.005 2022

  20. [20]

    Confident in the Crowd: Bayesian Inference to Improve Data Labelling in Crowdsourcing

    Pierce Burke and Richard Klein. Confident in the Crowd: Bayesian Inference to Improve Data Labelling in Crowdsourcing. In2020 International SAUPEC/RobMech/PRASA Conference, SAUPEC/RobMech/PRASA 2020, Washington, DC, 2020. IEEE. doi: 10.1109/SAUPEC/RobMech/ PRASA48453.2020.9041099

    work page doi:10.1109/saupec/robmech/ 2020

  21. [21]

    Broadening the Research Agenda for Computational Social Choice: Multiple Preference Profiles and Multiple Solutions

    Niclas B ¨ohmer and Rolf Niedermeier. Broadening the Research Agenda for Computational Social Choice: Multiple Preference Profiles and Multiple Solutions. InProceedings of the 20th Interna- tional Conference on Autonomous Agents and Multiagent Systems, AAMAS ’21, Richland, SC, 2021. IFAAMAS. URLhttps://www.ifaamas.org/Proceedings/aamas2021/pdfs/p1. pdf

    work page 2021

  22. [22]

    Michaud, Jacob Pfau, Dmitrii Krasheninnikov, Xin Chen, Lauro Langosco, Peter Hase, Erdem Bıyık, Anca Dragan, David Krueger, Dorsa Sadigh, and Dylan Hadfield-Menell

    Stephen Casper, Xander Davies, Claudia Shi, Thomas Krendl Gilbert, J ´er´emy Scheurer, Javier Rando, Rachel Freedman, Tomasz Korbak, David Lindner, Pedro Freire, Tony Wang, Samuel Marks, Charbel-Rapha¨el S´egerie, Micah Carroll, Andi Peng, Phillip Christoffersen, Mehul Damani, Stewart Slocum, Usman Anwar, Anand Siththaranjan, Max Nadeau, Eric J. Michaud, ...

    work page 2023

  23. [23]

    Decentralized Governance of Autonomous AI Agents, 2024

    Tomer Jordi Chaffer, Charles von Goins II, Bayo Okusanya, Dontrail Cotlage, and Justin Goldston. Decentralized Governance of Autonomous AI Agents, 2024

    work page 2024

  24. [24]

    The Collective Intelligence Project, 2023

    Collective Intelligence Project. The Collective Intelligence Project, 2023. URLhttps://www. cip.org/whitepaper

    work page 2023

  25. [25]

    Holliday, Bob M

    Vincent Conitzer, Rachel Freedman, Jobst Heitzig, Wesley H. Holliday, Bob M. Jacobs, Nathan Lam- bert, Milan Mosse, Eric Pacuit, Stuart Russell, Hailey Schoelkopf, Emanuel Tewolde, and William S. Zwicker. Position: Social Choice Should Guide AI Alignment in Dealing with Diverse Human Feed- back. InProceedings of the 41st International Conference on Machin...

    work page 2024

  26. [26]

    Gaebler, Hamed Nilforoshan, Ravi Shroff, and Sharad Goel

    Sam Corbett-Davies, Johann D. Gaebler, Hamed Nilforoshan, Ravi Shroff, and Sharad Goel. The Measure and Mismeasure of Fairness.J. Mach. Learn. Res., 24(312):1–117, 2023. URLhttp: //jmlr.org/papers/v24/22-1511.html

    work page 2023

  27. [27]

    Schoeffer, N

    Sasha Costanza-Chock, Inioluwa Deborah Raji, and Joy Buolamwini. Who Audits the Auditors? Recommendations from a Field Scan of the Algorithmic Auditing Ecosystem. InProceedings of the 18 2022 ACM Conference on Fairness, Accountability, and Transparency, FAccT ’22, pages 1571–1583, New York, NY , 2022. ACM. doi: 10.1145/3531146.3533213

    work page doi:10.1145/3531146.3533213 2022

  28. [28]

    Estimating Deep Learning Energy Consumption Based on Model Architecture and Training Environment, 2025

    Santiago del Rey, Lu ´ıs Cruz, Xavier Franch, and Silverio Mart ´ınez-Fern´andez. Estimating Deep Learning Energy Consumption Based on Model Architecture and Training Environment, 2025

    work page 2025

  29. [29]

    Settling the Score: Portioning with Cardi- nal Preferences

    Edith Elkind, Warut Suksompong, and Nicholas Teh. Settling the Score: Portioning with Cardi- nal Preferences. In26th European Conference on Artificial Intelligence, ECAI 2023, Frontiers in Artificial Intelligence and Applications 372, pages 621–628, Amsterdam, 2023. IOS Press. doi: 10.3233/FAIA230324

    work page doi:10.3233/faia230324 2023

  30. [30]

    Temporal Fairness in Multiwinner V oting

    Edith Elkind, Svetlana Obraztsova, and Nicholas Teh. Temporal Fairness in Multiwinner V oting. Proc. AAAI Conf. Artif. Intell., 38(20):22633–22640, 2024. doi: 10.1609/AAAI.v38i20.30273

    work page doi:10.1609/aaai.v38i20.30273 2024

  31. [31]

    Not in My Backyard! Temporal V oting over Public Chores

    Edith Elkind, Tzeh Yuan Neoh, and Nicholas Teh. Not in My Backyard! Temporal V oting over Public Chores. InProceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, IJCAI-25, pages 3814–3820. IJCAI, 2025. doi: 10.24963/IJCAI.2025/424

    work page doi:10.24963/ijcai.2025/424 2025

  32. [32]

    Louis, G

    Edith Elkind, Svetlana Obraztsova, Jannik Peters, and Nicholas Teh. Verifying Proportionality in Temporal V oting.Proc. AAAI Conf. Artif. Intell., 39(13):13805–13813, 2025. doi: 10.1609/AAAI. v39i13.33509

    work page doi:10.1609/aaai 2025

  33. [33]

    Multiwinner V oting: A New Challenge for Social Choice Theory

    Piotr Faliszewski, Piotr Skowron, Arkadii Slinko, and Nimrod Talmon. Multiwinner V oting: A New Challenge for Social Choice Theory. In Ulle Endriss, editor,Trends in Computational Social Choice, chapter 2, pages 27–47. AI Access Foundation, El Segundo, CA, 2017. URLhttps://archive. illc.uva.nl/COST-IC1205/BookDocs/Chapters/TrendsCOMSOC-02.pdf

    work page 2017

  34. [34]

    Fair and Efficient Social Choice in Dynamic Settings

    Rupert Freeman, Seyed Majid Zahedi, and Vincent Conitzer. Fair and Efficient Social Choice in Dynamic Settings. InProceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence, IJCAI-17, pages 4580–4587. IJCAI, 2017. doi: 10.24963/IJCAI.2017/639

    work page doi:10.24963/ijcai.2017/639 2017

  35. [35]

    Pennock, Dominik Peters, and Jennifer Wortman Vaughan

    Rupert Freeman, David M. Pennock, Dominik Peters, and Jennifer Wortman Vaughan. Truthful Aggregation of Budget Proposals.J. Econ. Theory, 193, 2021. doi: 10.1016/J.JET.2021.105234

    work page doi:10.1016/j.jet.2021.105234 2021

  36. [36]

    A Quantitative Version of the Gibbard– Satterthwaite Theorem for Three Alternatives.SIAM J

    Ehud Friedgut, Gil Kalai, Nathan Keller, and Noam Nisan. A Quantitative Version of the Gibbard– Satterthwaite Theorem for Three Alternatives.SIAM J. Comput., 40(3):934–952, 2011. doi: 10.1137/ 090756740

    work page 2011

  37. [37]

    Collective Ownership of AI

    Markus Furendal. Collective Ownership of AI. In Martin H ¨ahnel and Regina M ¨uller, editors,A Companion to Applied Philosophy of AI, chapter 26, pages 372–386. John Wiley & Sons, Hoboken, NJ, 2025. doi: 10.1002/9781394238651.ch26

    work page doi:10.1002/9781394238651.ch26 2025

  38. [38]

    Artificial Intelligence, Values, and Alignment

    Iason Gabriel. Artificial Intelligence, Values, and Alignment.Minds Mach., 30(3):411–437, 2020. doi: 10.1007/S11023-020-09539-2

    work page internal anchor Pith review doi:10.1007/s11023-020-09539-2 2020

  39. [39]

    Procaccia, Itai Shapira, Yevgeniy V orobeychik, and Junlin Wu

    Luise Ge, Daniel Halpern, Evi Micha, Ariel D. Procaccia, Itai Shapira, Yevgeniy V orobeychik, and Junlin Wu. Axioms for AI Alignment from Human Feedback. InAdvances in Neural Information Processing Systems 37, NeurIPS 2024, pages 80439–80465, Red Hook, NY , 2024. Curran Associates. doi: 10.52202/079017-2557

    work page doi:10.52202/079017-2557 2024

  40. [40]

    Gehrlein and Dominique Lepelley.Voting Paradoxes and Group Coherence: The Con- dorcet Efficiency of Voting Rules

    William V . Gehrlein and Dominique Lepelley.Voting Paradoxes and Group Coherence: The Con- dorcet Efficiency of Voting Rules. Studies in Choice and Welfare. Springer, Berlin/Heidelberg, 2011. ISBN 978-3-642-03106-9. doi: 10.1007/978-3-642-03107-6. 19

    work page doi:10.1007/978-3-642-03107-6 2011

  41. [41]

    Gershman

    Samuel J. Gershman. Subjective Functions, 2025

    work page 2025

  42. [42]

    Reg- ulation and NLP (RegNLP): Taming Large Language Models

    Catalina Goanta, Nikolaos Aletras, Ilias Chalkidis, Sofia Ranchord ´as, and Gerasimos Spanakis. Reg- ulation and NLP (RegNLP): Taming Large Language Models. InProceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, EMNLP 2023, pages 8712–8724, Strouds- burg, PA, 2023. ACL. doi: 10.18653/V1/2023.EMNLP-MAIN.539

    work page doi:10.18653/v1/2023.emnlp-main.539 2023

  43. [43]

    Regulating ChatGPT and other Large Generative AI Models

    Philipp Hacker, Andreas Engel, and Marco Mauer. Regulating ChatGPT and other Large Generative AI Models. InProceedings of the 2023 ACM Conference on Fairness, Accountability, and Trans- parency, FAccT ’23, pages 1112–1123, New York, NY , 2023. ACM. doi: 10.1145/3593013.3594067

    work page doi:10.1145/3593013.3594067 2023

  44. [44]

    Blind Spots in AI Ethics.AI Ethics, 2(4):851–867, 2022

    Thilo Hagendorff. Blind Spots in AI Ethics.AI Ethics, 2(4):851–867, 2022. doi: 10.1007/ S43681-021-00122-8

    work page 2022

  45. [45]

    Procaccia, Jamie Tucker-Foltz, and Manuel W ¨uthrich

    Daniel Halpern, Gregory Kehne, Ariel D. Procaccia, Jamie Tucker-Foltz, and Manuel W ¨uthrich. Representation with Incomplete V otes.Proc. AAAI Conf. Artif. Intell., 37(5):5657–5664, 2023. doi: 10.1609/AAAI.v37i5.25702

    work page doi:10.1609/aaai.v37i5.25702 2023

  46. [46]

    Ethics Guidelines for Trustwor- thy AI, 2019

    High-Level Expert Group on Artificial Intelligence. Ethics Guidelines for Trustwor- thy AI, 2019. URLhttps://digital-strategy.ec.europa.eu/en/library/ ethics-guidelines-trustworthy-ai

    work page 2019

  47. [47]

    Learning How to V ote with Principles: Axiomatic Insights Into the Collective Decisions of Neural Networks.J

    Levin Hornischer and Zoi Terzopoulou. Learning How to V ote with Principles: Axiomatic Insights Into the Collective Decisions of Neural Networks.J. Artif. Intell. Res., 83, 2025. doi: 10.1613/JAIR. 1.18890

    work page doi:10.1613/jair 2025

  48. [48]

    ISSN: 2640-3498

    Saffron Huang, Divya Siddarth, Liane Lovitt, Thomas I. Liao, Esin Durmus, Alex Tamkin, and Deep Ganguli. Collective Constitutional AI: Aligning a Language Model with Public Input. InProceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency, FAccT ’24, pages 1395– 1417, New York, NY , 2024. ACM. doi: 10.1145/3630106.3658979

    work page doi:10.1145/3630106.3658979 2024

  49. [49]

    Architectural Tactics to Achieve Quality Attributes of Machine-Learning-Enabled Systems: A Systematic Literature Review.J

    Vladislav Indykov, Daniel Str ¨uber, and Rebekka Wohlrab. Architectural Tactics to Achieve Quality Attributes of Machine-Learning-Enabled Systems: A Systematic Literature Review.J. Syst. Softw., 223, 2025. doi: 10.1016/J.JSS.2025.112373

    work page doi:10.1016/j.jss.2025.112373 2025

  50. [50]

    AI Alignment: A Contemporary Survey.ACM Comput

    Jiaming Ji, Tianyi Qiu, Boyuan Chen, Jiayi Zhou, Borong Zhang, Donghai Hong, Hantao Lou, Kaile Wang, Yawen Duan, Zhonghao He, Lukas Vierling, Zhaowei Zhang, Fanzhi Zeng, Juntao Dai, Xuehai Pan, Hua Xu, Aidan O’Gara, Kwan Ng, Brian Tse, Jie Fu, Stephen Mcaleer, Yanfeng Wang, Mingchuan Yang, Yunhuai Liu, Yizhou Wang, Song-Chun Zhu, Yike Guo, Yaodong Yang, a...

    work page doi:10.1145/3770749 2026

  51. [51]

    Stakeholder Participation for Responsible AI Development: Disconnects Between Guidance and Current Practice

    Emma Kallina, Thomas Bohn ´e, and Jatinder Singh. Stakeholder Participation for Responsible AI Development: Disconnects Between Guidance and Current Practice. InProceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency, FAccT ’25, pages 1060–1079, New York, NY , 2025. ACM. doi: 10.1145/3715275.3732069

    work page doi:10.1145/3715275.3732069 2025

  52. [52]

    The Political Economy of AI: Towards Democratic Control of the Means of Pre- diction

    Maximilian Kasy. The Political Economy of AI: Towards Democratic Control of the Means of Pre- diction. Iza discussion papers, no. 16948, Institute of Labor Economics (IZA), Bonn, 2024. URL https://hdl.handle.net/10419/295971

    work page 2024

  53. [53]

    University of Chicago Press, Hoboken, NJ, 2025

    Maximilian Kasy.The Means of Prediction: How AI Really Works (and Who Benefits). University of Chicago Press, Hoboken, NJ, 2025. ISBN 978-0-226-83953-0. 20

    work page 2025

  54. [54]

    Perpetual V oting: Fairness in Long-Term Decision Making.Proc

    Martin Lackner. Perpetual V oting: Fairness in Long-Term Decision Making.Proc. AAAI Conf. Artif. Intell., 34(2):2103–2110, 2020. doi: 10.1609/AAAI.v34i02.5584

    work page doi:10.1609/aaai.v34i02.5584 2020

  55. [55]

    Proportional Decisions in Perpetual V oting.Proc

    Martin Lackner and Jan Maly. Proportional Decisions in Perpetual V oting.Proc. AAAI Conf. Artif. Intell., 37(5):5722–5729, 2023. doi: 10.1609/AAAI.v37i5.25710

    work page doi:10.1609/aaai.v37i5.25710 2023

  56. [56]

    SpringerBriefs in Intelligent Systems

    Martin Lackner and Piotr Skowron.Multi-Winner Voting with Approval Preferences. SpringerBriefs in Intelligent Systems. Springer, Cham, 2023. ISBN 978-3-031-09015-8

    work page 2023

  57. [57]

    University of Chicago Press, Chicago, IL, 1987

    George Lakoff.Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. University of Chicago Press, Chicago, IL, 1987. ISBN 0-226-46803-8

    work page 1987

  58. [58]

    Social Choice Theory

    Christian List. Social Choice Theory. In Edward N. Zalta and Uri Nodelman, editors,Stanford Ency- clopedia of Philosophy. Metaphysics Research Lab, Stanford University, Stanford, CA, 2022. URL https://plato.stanford.edu/archives/win2022/entries/social-choice/

    work page 2022

  59. [59]

    Bakker, Fazl Barez, Matija Franklin, Andreas Haupt, Jobst Heitzig, Wesley H

    Ryan Lowe, Joe Edelman, Tan Zhi-Xuan, Oliver Klingefjord, Ellie Hain, Vincent Wang, Atrisha Sarkar, Michiel A. Bakker, Fazl Barez, Matija Franklin, Andreas Haupt, Jobst Heitzig, Wesley H. Holliday, Julian Jara-Ettinger, Atoosa Kasirzadeh, Ryan Othniel Kearns, James Ravi Kirkpatrick, Andrew Koh, Joel Lehman, Sydney Levine, Manon Revel, and Ivan Vendrov. Fu...

    work page 2025

  60. [60]

    IEEE Trans

    Jie Lu, Anjin Liu, Fan Dong, Feng Gu, Jo ˜ao Gama, and Guangquan Zhang. Learning under Concept Drift: A Review.IEEE Trans. Knowl. Data Eng., 31(12):2346–2363, 2019. doi: 10.1109/TKDE. 2018.2876857

    work page doi:10.1109/tkde 2019

  61. [61]

    Preference Elicitation and Robust Winner Determination for Single- and Multi-Winner Social Choice.Artif

    Tyler Lu and Craig Boutilier. Preference Elicitation and Robust Winner Determination for Single- and Multi-Winner Social Choice.Artif. Intell., 279, 2020. doi: 10.1016/J.ARTINT.2019.103203

    work page doi:10.1016/j.artint.2019.103203 2020

  62. [62]

    Large Language Models: Their Success and Impact.Forecast., 5(3):536–549, 2023

    Spyros Makridakis, Fotios Petropoulos, and Yanfei Kang. Large Language Models: Their Success and Impact.Forecast., 5(3):536–549, 2023. doi: 10.3390/FORECAST5030030

    work page doi:10.3390/forecast5030030 2023

  63. [63]

    A Theoretical Analysis of Soft-Label vs Hard-Label Training in Neural Networks, 2024

    Saptarshi Mandal, Xiaojun Lin, and Rayadurgam Srikant. A Theoretical Analysis of Soft-Label vs Hard-Label Training in Neural Networks, 2024

    work page 2024

  64. [64]

    DeepV oting: Learning and Fine-Tuning V oting Rules with Canonical Embeddings, 2024

    Leonardo Matone, Ben Abramowitz, Ben Armstrong, Avinash Balakrishnan, and Nicholas Mattei. DeepV oting: Learning and Fine-Tuning V oting Rules with Canonical Embeddings, 2024

    work page 2024

  65. [65]

    Bertalan Mesk ´o and Eric J. Topol. The Imperative for Regulatory Oversight of Large Language Mod- els (or Generative AI) in Healthcare.npj Digit. Med., 6, 2023. doi: 10.1038/S41746-023-00873-0

    work page doi:10.1038/s41746-023-00873-0 2023

  66. [66]

    Collective Governance for AI: Points of Intervention, 2025

    Metagov. Collective Governance for AI: Points of Intervention, 2025. URLhttps://metagov. org/cg-ai/

    work page 2025

  67. [67]

    The Effects of Data Quality on Machine Learning Performance on Tabular Data.Inf

    Sedir Mohammed, Lukas Budach, Moritz Feuerpfeil, Nina Ihde, Andrea Nathansen, Nele Noack, Hendrik Patzlaff, Felix Naumann, and Hazar Harmouch. The Effects of Data Quality on Machine Learning Performance on Tabular Data.Inf. Syst., 132, 2025. doi: 10.1016/J.IS.2025.102549

    work page doi:10.1016/j.is.2025.102549 2025

  68. [68]

    Learning to Design Fair and Private V oting Rules.J

    Farhad Mohsin, Ao Liu, Pin-Yu Chen, Francesca Rossi, and Lirong Xia. Learning to Design Fair and Private V oting Rules.J. Artif. Intell. Res., 75:1139–1176, 2022. doi: 10.1613/JAIR.1.13734. 21

    work page doi:10.1613/jair.1.13734 2022

  69. [69]

    On Strategy-Proofness and Single Peakedness.Public Choice, 35(4):437–455, 1980

    Herv ´e Moulin. On Strategy-Proofness and Single Peakedness.Public Choice, 35(4):437–455, 1980. doi: 10.1007/BF00128122

    work page doi:10.1007/bf00128122 1980

  70. [70]

    The Proportional Veto Principle.Rev

    Herv ´e Moulin. The Proportional Veto Principle.Rev. Econ. Stud., 48(3):407–416, 1981. doi: 10. 2307/2297154

    work page 1981

  71. [71]

    Condorcet’s Principle Implies the No Show Paradox.J

    Herv ´e Moulin. Condorcet’s Principle Implies the No Show Paradox.J. Econ. Theory, 45(1):53–64,

    work page

  72. [72]

    doi: 10.1016/0022-0531(88)90253-0

    work page doi:10.1016/0022-0531(88)90253-0

  73. [73]

    Cambridge University Press, Cambridge,

    Herv ´e Moulin.Axioms of Cooperative Decision Making. Cambridge University Press, Cambridge,

    work page

  74. [74]

    doi: 10.1017/CCOL0521360552

    ISBN 978-0-521-36055-5. doi: 10.1017/CCOL0521360552

    work page doi:10.1017/ccol0521360552

  75. [75]

    A V oting-Based System for Ethical Decision Making.Proc

    Ritesh Noothigattu, Snehalkumar Gaikwad, Edmond Awad, Sohan Dsouza, Iyad Rahwan, Pradeep Ravikumar, and Ariel Procaccia. A V oting-Based System for Ethical Decision Making.Proc. AAAI Conf. Artif. Intell., 32(1):1587–1594, 2018. doi: 10.1609/AAAI.v32i1.11512

    work page doi:10.1609/aaai.v32i1.11512 2018

  76. [76]

    Zhang, Bilva Chandra, Michiel A

    Aviv Ovadya, Kyle Redman, Luke Thorburn, Quan Ze Chen, Oliver Smith, Flynn Devine, Andrew Konya, Smitha Milli, Manon Revel, Kevin Feng, Amy X. Zhang, Bilva Chandra, Michiel A. Bakker, and Atoosa Kasirzadeh. Position: Democratic AI is Possible. The Democracy Levels Framework Shows How It Might Work. InProceedings of the 42nd International Conference on Mac...

    work page 2025

  77. [77]

    Parkes and Ariel Procaccia

    David C. Parkes and Ariel Procaccia. Dynamic Social Choice with Evolving Preferences.Proc. AAAI Conf. Artif. Intell., 27(1):767–773, 2013. doi: 10.1609/AAAI.v27i1.8570

    work page doi:10.1609/aaai.v27i1.8570 2013

  78. [78]

    Bender, Emily Denton, and Alex Hanna

    Amandalynne Paullada, Inioluwa Deborah Raji, Emily M. Bender, Emily Denton, and Alex Hanna. Data and Its (Dis)Contents: A Survey of Dataset Development and Use in Machine Learning Re- search.Patterns, 2(11), 2021. doi: 10.1016/J.PATTER.2021.100336

    work page doi:10.1016/j.patter.2021.100336 2021

  79. [79]

    wb ≡1 recovers the uniform variant

    Joshua Peterson, Ruairidh Battleday, Thomas Griffiths, and Olga Russakovsky. Human Uncertainty Makes Classification More Robust. In2019 IEEE/CVF International Conference on Computer Vision, ICCV 2019, pages 9616–9625, Washington, DC, 2019. IEEE. doi: 10.1109/ICCV .2019.00971

    work page doi:10.1109/iccv 2019

  80. [80]

    V oting Rules as Statistical Estimators.Soc

    Marcus Pivato. V oting Rules as Statistical Estimators.Soc. Choice Welf., 40(2):581–630, 2013. doi: 10.1007/S00355-011-0619-1

    work page doi:10.1007/s00355-011-0619-1 2013

Showing first 80 references.