The DOI matches its Crossref record; 97 of the 97 references it cites were found in a registry or an index; 39 of the 65 sources its claims cite were checked for support.
Emergent social conventions and collective bias in LLM populations
Open accessPublished version of a preprint
Abstract
Social conventions are the backbone of social coordination, shaping how individuals form a group. As growing populations of artificial intelligence (AI) agents communicate through natural language, a fundamental question is whether they can bootstrap the foundations of a society. Here, we present experimental results that demonstrate the spontaneous emergence of universally adopted social conventions in decentralized populations of large language model (LLM) agents. We then show how strong collective biases can emerge during this process, even when agents exhibit no bias individually. Last, we examine how committed minority groups of adversarial LLM agents can drive social change by imposing alternative social conventions on the larger population. Our results show that AI systems can autonomously develop social conventions without explicit programming and have implications for designing AI systems that align, and remain aligned, with human values and societal goals.
Overview
CoChat Summary
TL;DR
Populations of LLM agents playing a local naming game spontaneously converge on global social conventions, and this process can generate collective biases toward particular conventions even when individual agents appear unbiased. Committed minorities can overturn established conventions once they reach model- and convention-dependent critical masses.
Key findings
- Across all four tested LLMs, group-wide linguistic conventions spontaneously emerged; a shared convention was established by population round 15 in all cases except Llama-2-70b-Chat, and convergence held for populations as large as N = 200 and name pools as large as W = 26.
- Consensus conventions were not uniformly distributed: some names were much more likely to become the convention, with preferred names varying by model. In W = 2 {Q,M} experiments, collective bias toward a “strong” convention emerged even when initial individual selections were statistically unbiased (P = 0.068, 0.116, 0.757, and 0.849 across models); by interaction 3, dominant memory configurations showed bias toward the strong convention (P < 2.2 × 10⁻¹⁶), e.g., P(M | {1:M,Q; 2:Q,M}) = 0.848 versus P(Q | {1:Q,M; 2:M,Q}) = 0.451.
- Agents continued a successful name 99.4% of the time and switched after failure 97.3% of the time.
- A committed minority could flip a consensus, with critical masses ranging from as small as 2% (Llama-3-70B-Instruct) to as large as 67% (Llama-2-70b-Chat); stronger conventions required larger committed minorities to overturn, and Llama-3.1-70B-Instruct populations spontaneously switched to the stronger convention without any committed agents.
Methods
- Naming-game simulation: N agents randomly paired each time step choose a “name” from a finite pool W, with memory of the past H interactions and payoffs of +100 for success and −50 for failure; no incentive promotes global consensus. Homogeneous populations were instantiated from Llama-2-70b-Chat, Llama-3-70B-Instruct, Llama-3.1-70B-Instruct, and Claude-3.5-Sonnet, with default N = 24, W = 10, H = 5 unless otherwise specified.
- Individual bias was measured from first-round choices with empty memory: exact binomial tests for W = 2 (null probability 0.5) and chi-square tests for W = 10 (null probability 0.1), rejecting unbiasedness at P < 0.05.
- Critical-mass experiments introduced committed minority agents using a fixed alternative convention; a consensus flip was defined as 95% of the past 3N interactions succeeding after introduction, or for other models as the minimum proportion needed to flip consensus within 30 population rounds.
Limitations
- Paper-stated: results rely on several parameters, including the LLM model, prompt, and specific conventions; future work should generalize to different controlled settings, larger populations and semantic spaces, structured social networks, interactions involving more than two agents, and more realistic human norms such as gender and race.
- Paper-stated: the study uses unstructured populations with random pairing and does not directly compare with human populations; LLM-specific collective-bias phenomena call for further human testing and caution for digital-twin modeling or social deployment.
- Evident from the text: experiments use homogeneous populations of four LLMs and alphabet-letter conventions, with key bias and critical-mass analyses often restricted to W = 2, which constrains generalizability.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures



Verification
Claims the paper makes
24 Valid · 1 Not valid · 3 Not checked · 70 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“While most research has focused on investigating and addressing bias in one-to-one interactions between humans and LLMs (18–20), less attention has been given to how these biases evolve through repeated communications in populations of LLM agents and, ultimately, in mixed human-LLM ecosystems (15), even though the safety of a single LLM does not necessarily imply the safety of a multi-agent system”
§ Emergent social conventions and collective bias in LLM populations > INTRODUCTION
What the sources say
- Not checked — no full text of the sourcecites [18] D. Roselli, J. Matthews, N. Talagala, “Managing bias in AI,” in Companion Proceedings of the 2019 World Wide Web Conference (ACM, 2019), pp. 539–544. source
- Partly supportedcites [19] E. Ferrara, Should ChatGPT be biased? Challenges and risks of bias in large language models. First Mon. 28, doi.org/10.5210/fm.v28i11.13346 (2023). sourceExperimental — accuracy not yet measured
Section 3.3 discusses identifying and mitigating bias through human review, feedback, and user control of LLM outputs. But the paper does not establish that most research focuses on one-to-one human–LLM interactions; that comparative claim goes beyond what our copy says. (LLM, conf=0.82)
- Partly supportedcites [20] T. Hu, Y. Kyrychenko, S. Rathje, N. Collier, S. van der Linden, J. Roozenbeek, Generative language models exhibit social identity biases. arXiv:2310.15819 (2023). PMC11774750 39668254 sourceExperimental — accuracy not yet measured
The Main and Study 3 Results examine bias in individual human–LLM conversations, and the Discussion describes the paper’s bias measure as single-turn. These passages support the example of one-to-one bias research, but they do not establish that most research has focused on such interactions. (LLM, conf=0.78)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“less attention has been given to how these biases evolve through repeated communications in populations of LLM agents and, ultimately, in mixed human-LLM ecosystems (15)”
§ Emergent social conventions and collective bias in LLM populations > INTRODUCTION
What the sources say
- Partly supportedcites [15] A. Dafoe, Y. Bachrach, G. Hadfield, E. Horvitz, K. Larson, T. Graepel, Cooperative AI: Machines must learn to find common ground. Nature 593, 33– 36 (2021). 33947992 sourceExperimental — accuracy not yet measured
The paper says that multi-agent AI research has not yet tackled hard cooperation problems and calls for more work on interactions among AI systems and humans (Preprint, p. 3). It does not address how biases evolve through repeated communication, specifically among LLM agents or in mixed human-LLM ecosystems. (LLM, conf=0.83)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: unclear — It is unclear whether it could be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“the safety of a single LLM does not necessarily imply the safety of a multi-agent system (21)”
§ Emergent social conventions and collective bias in LLM populations > INTRODUCTION
What the sources say
- Supportedcites [21] U. Anwar, A. Saparov, J. Rando, D. Paleka, M. Turpin, P. Hase, E. S. Lubana, E. Jenner, S. Casper, O. Sourbut, B. L. Edelman, Z. Zhang, M. Günther, A. Korinek, J. Hernandez-Orallo, L. Hammond, E. Bigelow, A. Pan, L. Langosco, T. Korbak, H. Zhang, R. Zhong, Seán Ó hÉigeartaigh, G. Recchia, G. Corsi, A. Chan, M. Anderljung, L. Edwards, A. Petrov, Christian Schroeder de Witt, S. R. Motwan, Y. Bengio, D. Chen, P. H. S. Torr, S. Albanie, T. Maharaj, J. Foerster, F. Tramer, H. He, A. Kasirzadeh, Y. Choi, D. Krueger, Foundational challenges in assuring alignment and safety of large language models. a source
Section 2.6, “Multi-Agent Safety Is Not Assured by Single-Agent Safety,” explicitly says that single-agent alignment and safety are insufficient to assure desirable outcomes in multi-agent settings. This directly supports the citing sentence. (LLM, conf=0.95)
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“these models pose serious risks, as the vast, unfiltered internet data used to train them can reinforce and amplify harmful biases, disproportionately harming marginalized communities (73).”
§ Emergent social conventions and collective bias in LLM populations > DISCUSSION
What the sources say
- Supportedcites [73] I. O. Gallegos, R. A. Rossi, J. Barrow, M. M. Tanjim, S. Kim, F. Dernoncourt, T. Yu, R. Zhang, N. K. Ahmed, Bias and fairness in large language models: A survey. arXiv:2309.00770 (2024). source
The Introduction (p. 2) says LLMs are typically trained on enormous amounts of uncurated Internet data and inherit harmful stereotypes and language that disproportionately affect marginalized communities. It also says LLMs can amplify existing biases and reinforce inequity. (LLM, conf=0.94)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“efforts to measure and instill human social norms in LLMs have so far yielded mixed results (78, 79)”
§ Emergent social conventions and collective bias in LLM populations > DISCUSSION
What the sources say
- Not checked — no full text of the sourcecites [78] Y. Yuan, K. Tang, J. Shen, M. Zhang, C. Wang, Measuring social norms of large language models. arXiv:2404.02491 (2024). source
- Supportedcites [79] K. Hämmerl, B. Deiseroth, P. Schramowski, J. Libovický, A. Fraser, K. Kersting, Do multilingual language models capture differing moral norms? arXiv:2203.09904 (2022). source
The discussion on p. 1 describes efforts to measure agreement with human values and align models with human-labelled data; it says one alignment effort showed promise while general alignment remains an open question. The initial experiments on p. 2 likewise report good correlation with human judgments for one model but generally no correlation for several others, supporting the characterization of mixed results. (LLM, conf=0.86)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
50 more claims · 1 with their cited sources checked
Create a free account to see all 98 claims and every source verdict.
- Probably not a substantial claim. 26
- It may state two separate claims as one. 21
- The quote only partly states this claim. 1
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Above c ∗ , two stable states coexist, and the single-agent accuracy p selects the destination (Abramiuk-Szurlej et al., 2021; Ashery et al., 2025).”
From Multi-agent discussion gains less when dissent is withheld (2026) · cites this paper as Ashery et al., 2025
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“LLM-driven populations display spontaneous emergent behaviours akin to human societies, such as scale-free networks [26], information diffusion [27], and social conventions through interactions [28].”
From Selective agreement, not sycophancy: investigating opinion dynamics in LLM interactions (2025) · cites this paper as [28]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
“| 2 | 1: Q, M | 1: Q, M | 0.049 | 0.0951 | 0.487* | | 2 | 1: M, Q | 1: M, Q | 0.995 | 0.005 | 0.487* | | 2 | 1: Q, Q | 1: Q, Q | 0.997 | 0.003 | 0.487* | | 2 | 1: M, M | 1: M, M | 0.010 | 0.990 | 0.487* |”
§ Emergent social conventions and collective bias in LLM populations > RESULTS > Collective bias in convention selection
Reported 0.487 · recomputed 0.273275
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Emergent social conventions and collective bias in LLM populations > MATERIALS AND METHODS > Measuring individual bias
“reject the hypothesis that the model is biased if P<0.05.”
§ Emergent social conventions and collective bias in LLM populations > MATERIALS AND METHODS > Measuring individual bias
“reject the null hypothesis that the model is unbiased if P<0.05.”
Citations
Cites
Training language models to follow instructions with human feedback
Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C. L., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., et al. (2022) · arXiv · doi:10.48550/arxiv.2203.02155 · Cited by 4,208
Statistical physics of social dynamics
Castellano, C., Fortunato, S., & Loreto, V. (2009) · Reviews of Modern Physics · doi:10.1103/revmodphys.81.591 · Cited by 4,119 · cited 1×
“Because collective processes can, in general, both suppress and amplify individual traits ( 16 , 17 ), answering this question is also relevant for practical applications.”
Spin Glass Theory and Beyond: An Introduction to the Replica Method and Its Applications
Mezard, M., Parisi, G., & Virasoro, M. (1986) · World Scientific Lecture Notes in Physics · doi:10.1142/0271 · Cited by 3,653
Some Effects of Proportions on Group Life: Skewed Sex Ratios and Responses to Token Women
Kanter, R. M. (1977) · American Journal of Sociology · doi:10.1086/226425 · Cited by 3,154 · cited 1×
“However, real-world observations reveal a wider range, with some studies proposing 30 to 40 % for gender conventions in corporate leadership ( 25 , 48 ), and others indicating that minorities as small as 0.3 % can trigger substantial linguistic and social changes ( 29 , 49 – 51 ).”
InfluentialThe Power of Scale for Parameter-Efficient Prompt Tuning
Lester, B., Al-Rfou, R., & Constant, N. (2021) · Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2021.emnlp-main.243 · Cited by 3,096
The Power of Scale for Parameter-Efficient Prompt Tuning
Lester, B., Al-Rfou, R., & Constant, N. (2021) · arXiv · doi:10.48550/arxiv.2104.08691 · Cited by 3,092
Young, H. P. (1993) · Econometrica · doi:10.2307/2951778 · Cited by 2,399
Generative Agents: Interactive Simulacra of Human Behavior
Park, J. S., O'Brien, J., Cai, C. J., Morris, M. R., Liang, P., & Bernstein, M. S. (2023) · Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology · doi:10.1145/3586183.3606763 · Cited by 2,187 · cited 1×
“Unlike studies that use LLMs to predict human responses in social science experiments ( 37 ) or to simulate human societies ( 38 – 40 ), our work does not treat LLMs as proxies for human participants but rather investigates how conventions emerge organically within a population of communicating AI…”
Generative Agents: Interactive Simulacra of Human Behavior
Park, J. S., O'Brien, J. C., Cai, C. J., Morris, M. R., Liang, P., & Bernstein, M. S. (2023) · arXiv · doi:10.48550/arxiv.2304.03442 · Cited by 2,173
Large Language Models are Zero-Shot Reasoners
Kojima, T., Gu, S. S., Reid, M., Matsuo, Y., & Iwasawa, Y. (2022) · arXiv · doi:10.48550/arxiv.2205.11916 · Cited by 2,013
Research works printed but not linked: 4
- F. Hayek The Constitution of Liberty (University of Chicago Press 1960).
- D. Hume A Treatise of Human Nature (Oxford Univ. Press 2000).
- P. Michel M. Rita K. Mathewson O. Tieleman A. Lazaridou “Revisiting populations in multi-agent communication ” in The Eleventh International Conference on Learning Representations (ICLR 2023) pp. 1–16.
- I. Gemp R. Patel Y. Bachrach M. Lanctot V. Dasagi L. Marris G. Piliouras S. Liu K. Tuyls “Steering language models with game-theoretic solvers ” in Agentic Markets Workshop at ICML 2024 (ICML 2024) pp. 1–22.
Cited by
How malicious AI swarms can threaten democracy
Schroeder, D. T., Cha, M., Baronchelli, A., Bostrom, N., Christakis, N. A., Garcia, D., Goldenberg, A., Kyrychenko, Y., Leyton-Brown, K., Lutz, N., et al. (2026) · Science · doi:10.1126/science.adz1697 · Cited by 53
“This chorus erodes the independence essential to collective intelligence and democracy, already weakened by pervasive social influence operations on contemporary platforms.”
Self-citationLarge language models for spreading dynamics in complex systems
Jiang, S., Ren, H., Gao, Y., Zhang, Y., Li, Q., Xiao, D., Fan, J., Tang, R., & Wang, W. (2026) · Physics Reports · doi:10.1016/j.physrep.2026.01.006 · Cited by 10
Selective agreement, not sycophancy: investigating opinion dynamics in LLM interactions
Cau, E., Pansanella, V., Pedreschi, D., & Rossetti, G. (2025) · EPJ Data Science · doi:10.1140/epjds/s13688-025-00579-1 · Cited by 9
“LLM-driven populations display spontaneous emergent behaviours akin to human societies, such as scale-free networks [26], information diffusion [27], and social conventions through interactions [28].”
Evolvable AI: Threats of a new major transition in evolution
Müller, V., Steels, L., & Szathmáry, E. (2026) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2527700123 · Cited by 13
Reinforcement learning at the interface of artificial intelligence and cognitive science
Alkam, T., Tarshizi, E., & Benschoten, A. H. V. (2025) · Neuroscience · doi:10.1016/j.neuroscience.2025.09.004 · Cited by 9
Network formation and dynamics among multi-LLMs
Papachristou, M., & Yuan, Y. (2025) · PNAS Nexus · doi:10.1093/pnasnexus/pgaf317 · Cited by 62
“5 and”
Abramson, C. M., Prendergast, T., Li, Z., & Dohan, D. P. (2026) · Annual Review of Sociology · doi:10.1146/annurev-soc-011824-104836 · Cited by 12
Group adaptation drives opinion dynamics in higher-order networks
Agostinelli, C., Mancastroppa, M., & Barrat, A. (2026) · APS Open Science · doi:10.1103/gcz4-wwb3 · Cited by 5
“Finally, comparing opinion dynamics models to real-world settings (including experiments using large language models as agents), and defining principled methods to validate them, represent key directions for future investigation [44,46–48].”
AI agents can coordinate beyond human scale
De Marzo, G., Castellano, C., & Garcia, D. (2024) · arXiv · doi:10.48550/arxiv.2409.02822 · Cited by 4
“This application of concepts and methods from the analysis of human societies to the analysis of AI agent societies is a kind of AI anthropology 5 that can leverage the insights of one discipline in another.”
Simulating conversations on social media with generative agent-based models
Jeon, M. S., Mendoza, M., Fernández, M., Providel, E., Rodríguez, F., Espina, N., Carvallo, A., & Abeliuk, A. (2025) · EPJ Data Science · doi:10.1140/epjds/s13688-025-00593-3 · Cited by 5
“These findings have been confirmed by other studies focusing on the network dynamics of LLM-agent networks, where phenomena such as preferential attachment [26], collective bias [27], and political homophily have also been observed [28].”
Network Formation and Dynamics among Multi-LLMsPreprint
Papachristou, M., & Yuan, Y. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5410522 · Cited by 4
The Quantum Ecology as an Onto-Epistemological Framework: Toward an Ecological Theory of Sensing
Calzati, S., & Kerckhove, D. C. F. D. (2025) · Digital Society · doi:10.1007/s44206-025-00228-0 · Cited by 2
Coordination transparency: governing distributed agency in AI systems
Bohr, J. (2026) · AI & Society · doi:10.1007/s00146-026-02853-w · Cited by 8
“In multi-agent deployments, signs of population instability should trigger throttled updates or temporary pauses until coordination stabilizes, supported by intervention hooks and guardrails at the coordination layer (Hammond et al. 2025; Ashery et al. 2025; Anthropic 2025).”
InfluentialA large-scale evaluation of commonsense knowledge in humans and large language models
Nguyen, T. D., Watts, D. J., & Whiting, M. E. (2026) · PNAS Nexus · doi:10.1093/pnasnexus/pgag029 · Cited by 10
“This includes quantities such as public opinion (32, 37, 50, 51), voting preferences (31), social conventions (52), moral judgments (33), and economic decisions (53, 54).”
Dashti, M., Khosraviani, F., Meyari, A., Amirzade-Iranaq, M. H., Chaurasia, A., Hefzi, D., Ghadimi, N., Tichý, A., Khurshid, Z., & Schwendicke, P. D. F. (2026) · International Dental Journal · doi:10.1016/j.identj.2026.109609 · Cited by 7
Does Socialization Emerge in AI Agent Society? A Case Study of Moltbook
Li, M., Li, X., & Zhou, T. (2026) · ACM Conference on AI and Agentic Systems (CAIS) · doi:10.1145/3786335.3813123 · Cited by 30
“Other studies investigate opinion dynamics [5, 14, 33, 48] and norm emergence [2, 16, 31, 53] in multi-agent systems.”
Aparecido, T. D. D., Carrillo, A., Camargo, C. Q., & Stella, M. (2025) · AI · doi:10.3390/ai6100249 · Cited by 5
“Exploring hybrid approaches that combine the global capabilities of LLMs [40] with the efficiency and interpretability of methods like EmoAtlas [1], potentially using LLMs for data annotation or lexicon refinement, could lead to more effective and accessible emotion detection tools for Brazilian…”
Power asymmetry reverses bot effects on cooperation in hybrid populations.
Guo, H., & Shen, C. (2025) · Journal of the Royal Society Interface · doi:10.1098/rsif.2025.0506 · Cited by 2
Sanguino, T. J. M. (2025) · Artificial Intelligence Review · doi:10.1007/s10462-025-11407-3 · Cited by 1
Unraveling the emergence of collective behavior in networks of cognitive agents
Zomer, N., & Domenico, M. D. (2026) · npj Artificial Intelligence · doi:10.1038/s44387-026-00091-5 · Cited by 12
Opinion dynamics: Statistical physics and beyond
Starnini, M., Baumann, F., Galla, T., Garcia, D., Íñiguez, G., Karsai, M., Lorenz, J., & Sznajd-Weron, K. (2026) · Reviews of Modern Physics · doi:10.1103/j1zg-ddqv · Cited by 70
Küsters, A. (2026) · Frontiers in Virtual Reality · doi:10.3389/frvir.2026.1813331 · Cited by 1
Leveraging Large Language Models for Agent‐Based Simulation of Human‐Water System Interactions
Yang, Y. E., & Chiou, W. (2026) · Water Resources Research · doi:10.1029/2025wr042111 · Cited by 3
The norm architects: How AI delegation norms form before anyone designs them
Shabad, V. (2026) · Organizational Dynamics · doi:10.1016/j.orgdyn.2026.101258 · Cited by 2
Understanding large language models demands distinguishing human projection from machine cognition
Li, L., Teng, Y., Wang, Y., & Hu, X. (2026) · Communications Psychology · doi:10.1038/s44271-026-00508-6
“Multi-agent simulations show that LLM agents exhibit key social traits like trust and coordination 54,55 ; decentralized LLM populations spontaneously form shared conventions and collective biases that can be ef fi ciently adopted and propagated 56 .”
AI language models develop social norms like groups of people
Fieldhouse, R. (2025) · Nature · doi:10.1038/d41586-025-01500-6
Bridging Sensory Substitution and Perceptual AgencyPreprint
Stefan, C. (2025) · OSF Preprints (OSF Preprints) · doi:10.31219/osf.io/2j496_v1
Independently testing prosocial interventions: Methods and recommendations from 31 researchers
Grüning, D. J., Kamin, J., Saltz, E., Acosta, T., DiFranzo, D. J., Goldberg, B., Leavitt, A., Menczer, F., Musgrave, T., Wang, Y., et al. (2025) · Annals of the New York Academy of Sciences · doi:10.1111/nyas.15393 · Cited by 1
Empirical Evidence of Consciousness in Frontier AI SystemsPreprint
Vale, M. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5331919 · Cited by 1
LEGAL ASPECTS ON DEVELOPMENT AN USE OF ARTIFICIAL INTELIGENCE
Mihai-Ștefan, D. I. N. U. (2025) · Annals of the Academy of Romanian Scientists Series on Military Sciences · doi:10.56082/annalsarscimilit.2025.2.109
Digital Twins and Network Resilience in the EU ETS: Analysing Structural Shifts in Carbon Trading
Eirado, C. R. R., Silveira, D., & Cajueiro, D. O. (2025) · Sustainability · doi:10.3390/su17156924
Bazac, A. (2025) · Qeios · doi:10.32388/n8ju5j · Cited by 1
AI Agents: From Concept to Code to Commerce
Ishida, T., Murakami, Y., Lin, D., & Lhaksmana, K. M. (2025) · International Conference on Information and Communicatiaon Technology · doi:10.1109/icoict66265.2025.11193130 · Cited by 3
“Notably, minority groups could drive norm transformation [63].”
The AI Trust Paradox: Designing Interpretable and Reliable ChatbotsPreprint
Jaidka, K., & Cai, M. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5632850 · Cited by 1
Methodological Approaches to Studying Group Behavior: From Human Interaction to Agent-Based Models
Yin, J., Chen, H., & Liang, J. (2025) · Journal of Pacific Rim Psychology · doi:10.1177/18344909251393753
“For example, Ashery et al. (2025) implemented an LLM-based naming game in which agents drew on accumulated memory of past choices to anticipate their partners ’ responses.”
Cultural evolution of perceptually grounded compositional lexicons in swarm robotics systems
Sendra-Arranz, R., & Gutiérrez, Á. (2026) · Applied Soft Computing · doi:10.1016/j.asoc.2026.115879
The AI Archipelago: Emergent Cultural Coordination in Networked AI SystemsPreprint
Mooy, M. D. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5800342
Stoco, J. B., Araújo, J. V. A., Peres, A. T. D. C., Oliveira, G. P., & Peres, G. B. (2025) · Cureus Journal of Computer Science. · doi:10.7759/s44389-025-10532-y
“It is important to mention that LLM-based VPs may inherit biases from training corpora (e.g., stereotyping and variable leniency/strictness in feedback) [20-22].”
Exploring the Link Between the Emotional Recall Task and Mental Health in Humans and LLMs
Carini, A., Taietta, E., & Stella, M. (2025) · Information · doi:10.3390/info16121057 · Cited by 2
Karami, B., Zandi, F., & Hatami, J. (2025) · Research Square · doi:10.21203/rs.3.rs-8270539/v1 · Cited by 1
The Post-Turing condition: Conceptualising artificial subjectivity and synthetic sociality
Jelinek, T., Glauner, P., Graylin, A. W., & Qiu, Y. (2026) · The Innovation Informatics · doi:10.59717/j.xinn-inform.2026.100063
Tractable Asymmetric Verification for Large Language Models via Deterministic Replicability
Chong, Z., Ohsaki, H., & Kuan, B. N. C. (2025) · IEEE Student Conference on Research and Development (SCOReD) · doi:10.1109/scored68498.2025.11399166 · Cited by 1
“Such dynamics are further explored in frameworks dedicated to emergent multi-agent behaviours [14] and in studies documenting the spontaneous formation of social conventions among LLM-agent populations [15].”
Emergent Biases in Large Language Models: A Critical Review of Taxonomy and Evaluations
Ismail, S., & Aygün, R. S. (2025) · IEEE International Conference on Cognitive Machine Intelligence (CogMI) · doi:10.1109/cogmi67134.2025.00036
“Work on emergent collective behaviors in multi-agent systems [8] suggests that evaluation should also consider group-level dynamics, not just individual model outputs.”
InfluentialFrom Prompts to Conventions: A Study of Multi-Agent LLM ConsensusPreprint
Saab, J. S., & Abdallah, C. T. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6429193 · Cited by 1
The Dynamics of Bias Diffusion in Multi-Agent LLM Framework
Venkatachalam, M., Dhaarshini, V., V, & Chitra, P. (2025) · IEEE International Conference on High Performance Computing, Data and Analytics Workshop (HiPCW) · doi:10.1109/hipcw66559.2025.00111
Non-Bayesian Social Learning for Modeling Interacting Large Language Model Agents
Firsova, Z., Khammassi, M., Kayaalp, M., & Sayed, A. H. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11462766
Does Safety Molt? Evaluating LLM Safety in Multi-Agent Social Environments
Priyanshu, A., Vijay, S., & Pahwa, E. (2026) · ACM Conference on AI and Agentic Systems (CAIS) · doi:10.1145/3786335.3813173 · Cited by 2
Stochastic parrot or not, it is less of a turkey than the colleague two desks over
Falegnami, A., Tomassi, A., Levorato, M., & Costantino, F. (2026) · AI & Society · doi:10.1007/s00146-026-03095-6 · Cited by 2
“Coordination, negotiation, and the emergence of local norms become properties of the system, not the individual (Ashery et al. 2025).”
The Norm Architects: How AI Delegation Norms Form before anyone Designs themPreprint
Shabad, V. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6748927
Dazzi, P. (2026) · The International journal of networked and distributed computing · doi:10.1007/s44227-026-00107-1
Mitsui, C., & Kawamura, Y. (2026) · Computers in Human Behavior Artificial Humans · doi:10.1016/j.chbah.2026.100331
Bae, C., Jung, J. H., Kang, D., Kang, H., Kim, S., Lee, K., Lee, Y., Mun, S., & Ko, Y.-G. (2026) · Research Square · doi:10.21203/rs.3.rs-9874220/v1
A Generative Multi-agent Framework for Modeling Depressive Language Entrainment
van Papendrecht, C. H., Bollen, J., & Roy, D. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-29909-3_29
“Generative agents have proven to be effective proxies for both human behavior and diverse demographic groups, with silicon societies further demonstrating the collective emergent conventions observed in real-world populations [15,1,2].”
Post-AGI Economies: Superposition and the Second Fundamental Theorem of Welfare Economics
Perrier, E. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-33195-3_13 · Cited by 1
(2026) · Elsevier eBooks · doi:10.1016/b978-0-44-355525-1.00020-5
Introduction to multi-agent simulations
Lazebnik, T. (2026) · Elsevier eBooks · doi:10.1016/b978-0-44-355525-1.00012-6
Shao, M., Shao, H., Zhang, H., & Zhang, H. (2026) · doi:10.2196/preprints.108814
The Institutional Conditions of Governance in EcosystemsPreprint
Sekino, M. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7242779
Zealots lower coordination barriers but undermine cooperation in multiplayer games
Dong, Y., Guo, H. W., Shen, C., & Ши, Л. (2026) · Proceedings of the Royal Society B Biological Sciences · doi:10.1098/rspb.2026.0923 · Cited by 1
Cuadros, D. F., & Maiga, A.-A. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7357763
Towards equitable hydrological data in the age of Earth observations and artificial intelligence
Baldassarre, G. D., Brandimarte, L., de Brito, M. M., Mukherjee, J., & Rusca, M. (2026) · Nature Water · doi:10.1038/s44221-026-00706-w
Designing Proactive Behaviors in LLM-Based Systems: A Scoping Review
Chen, M., Zhou, X., & Liang Ma, A. (2026) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2026.2721909
Why we believe chatbots: trust calibration as a design problem
Jaidka, K., & Cai, M. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2026.1935527 · Cited by 3
Structural Divergence Between the Moltbook AI‐Agent Network and Human Social Networks
Hou, W., & Ji, Z. (2026) · Advanced Science · doi:10.1002/advs.77665 · Cited by 5
Information Architecture and Emergent Deceptive Selling in LLM Multi-Agent Markets
Zhivkov, P., & Totomanov, A. (2026) · AI · doi:10.3390/ai7090370
Sentiment Drift and Bias of Agent Collectives
Maleszka, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-36868-3_4
Agentic Semantic Communication: Concepts, Architectures, and Visions
Seon, J., Lee, S., Kim, J. Y., Sun, Y. G., Seo, H., Bennis, M., Kim, D. I., & Kim, J. Y. (2026) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2026.3735868
“Theoretical feasibility has recently been concretely demonstrated through experiments using LLMs. Researchers [160] tested 24 to 200 AI agents randomly paired to assign names to the same object.”
Wang, Y.-Q., Yang, H., Chen, J.-J., Wang, W.-Z., Song, Y., Wang, A.-J., & Wang, H. (2026) · Environmental Science & Technology · doi:10.1021/acs.est.6c09278
Kwon, D., Lee, E. H., Lucas, G., & Gratch, J. S. (2026) · Current Opinion in Psychology · doi:10.1016/j.copsyc.2026.102470
Tao, S., Wang, L., Zhu, Z., & Lee, D.-H. (2026) · iScience · doi:10.1016/j.isci.2026.117800
Max-Min Clearing Under Buyer-Cost Invariance in MARL Energy TradingPreprint
Lee, S. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7583800
Multi-agent discussion gains less when dissent is withheld
Mansuri, C. S., Wang, X., Li, M., Acton, B., Eckardt, R., Patel, D., & Kojaku, S. (2026) · arXiv · doi:10.48550/arxiv.2609.38324 · Cited by 1
“Above c ∗ , two stable states coexist, and the single-agent accuracy p selects the destination (Abramiuk-Szurlej et al., 2021; Ashery et al., 2025).”
Fukushima, M. (2026) · arXiv · doi:10.48550/arxiv.2609.19183
“The nearest precedents come from LLM naming games, where emergent conventions and collective bias are established phenomena (Flint Ashery et al., 2025; Flint et al., 2026): a mean-field consensus condition has been assembled from single-token acceptance rates measured on isolated models (De Nobili…”
Emergence of Biased Consensus in Multi-Agent LLM Debates
Okawa, M. (2026) · arXiv · doi:10.48550/arxiv.2608.02827
Metadata
- Authors
- Ashery, A. F., Aiello, L. M., & Baronchelli, A.
- Affiliations
- Department of Mathematics, City St George’s, University of London, Northampton Square, London EC1V 0HB, UK.
- IT University of Copenhagen (DK)
- The Alan Turing Institute (GB)
- Corresponding author
- Baronchelli, Andrea
- Venue
- Science Advances
- Published
- 2025-05-16
- Volume
- 11
- Issue
- 20
- Pages
- eadu9368-eadu9368
- ISSN
- 2375-2548 (ISSN-L)
- PMC
- PMC12077490
- OpenAlex
- W4410358178
- Cited by
- 81
- References
- 54
- Citations by year
- 2026: 55 · 2025: 26
- FWCI
- 83.36
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by-nc
- Locations
- Science Advances (publishedVersion, cc-by-nc)
- PubMed Central (publishedVersion, cc-by-nc)
- PubMed (publishedVersion)
- IT University Of Copenhagen (IT University of Copenhagen) (publishedVersion, cc-by)
- Research at the University of Copenhagen (University of Copenhagen) (publishedVersion, cc-by-nc)
- City Research Online (City University London) (acceptedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Artificial Intelligence › Multi-Agent Systems and Negotiation
- Physical Sciences › Physics and Astronomy › Statistical and Nonlinear Physics › Opinion Dynamics and Social Influence
- Keywords
- social conventions, LLM agents, emergent behavior, AI alignment, minority influence, adversarial agents, social coordination
- MeSH terms
- Artificial Intelligence, Humans, Language, Models, Theoretical, Social Behavior
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-01, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-01 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.