{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,23]],"date-time":"2026-09-23T00:02:29Z","timestamp":1790121749120,"version":"4.0.1"},"reference-count":58,"publisher":"American Association for the Advancement of Science (AAAS)","issue":"33","funder":[{"DOI":"10.13039\/100025178","name":"OpenAI","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100025178","id-type":"DOI","asserted-by":"publisher"}]},{"name":"European Union - NextGenerationEU","award":["Componente C2 Investimento 1.1"],"award-info":[{"award-number":["Componente C2 Investimento 1.1"]}]},{"name":"PRIN Project","award":["No. 20223W2JKJ \u201cWECARE\u201d"],"award-info":[{"award-number":["No. 20223W2JKJ \u201cWECARE\u201d"]}]},{"name":"Italian Ministry of University and Research (MUR), Piano Nazionale Di Ripresa e Resilienza (PNRR), Missione 4 \u201cIstruzione e Ricerca\u201d","award":["CUP B53D23003880006"],"award-info":[{"award-number":["CUP B53D23003880006"]}]}],"content-domain":{"domain":["www.science.org"],"crossmark-restriction":true},"short-container-title":["Sci. Adv."],"published-print":{"date-parts":[[2026,8,14]]},"abstract":"<jats:p>Large language models (LLMs) are increasingly deployed in collaborative tasks forming \u201cAI agent societies\u201d where agents interact and influence one another. Whether such groups can spontaneously coordinate without external influence, a hallmark of self-organized regulation in human societies, remains an open question. Here, we use principles from complexity and behavioral science to investigate coordination in AI agent groups through majority-following, a fundamental mechanism for spontaneous consensus formation. Using binary opinion dynamics experiments across multiple LLM architectures and group sizes, we find that agents exhibit majority-following characterized by a universal functional form with a single parameter, the \u201cmajority force.\u201d This majority force diminishes as group size increases, leading to a critical size beyond which coordination becomes unattainable. The critical group size grows rapidly with model capabilities and, for advanced LLMs, exceeds 1000 agents, larger than typical human informal groups. Our findings have implications for designing collaborative AI systems where coordination could be beneficial or pose safety threats.<\/jats:p>","DOI":"10.1126\/sciadv.aea6091","type":"journal-article","created":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T17:58:41Z","timestamp":1786730321000},"update-policy":"https:\/\/doi.org\/10.34133\/aaas_crossmark","source":"Crossref","is-referenced-by-count":2,"title":["AI agents can coordinate via majority-following beyond human scale"],"prefix":"10.1126","volume":"12","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-3127-5336","authenticated-orcid":true,"given":"Giordano","family":"De Marzo","sequence":"first","affiliation":[{"name":"University of Konstanz, Universitaetstrasse 10, Konstanz, 78457, Germany."},{"name":"Centro Ricerche Enrico Fermi, Piazza del Viminale 1, Rome, 00184, Italia."},{"name":"Complexity Science Hub, Josefstaedter Strasse 39, Vienna, 1080, Austria."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3773-3801","authenticated-orcid":true,"given":"Claudio","family":"Castellano","sequence":"additional","affiliation":[{"name":"Centro Ricerche Enrico Fermi, Piazza del Viminale 1, Rome, 00184, Italia."},{"name":"Istituto dei Sistemi Complessi (ISC-CNR), Via dei Taurini 19, Roma, 00185, Italia."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2820-9151","authenticated-orcid":true,"given":"David","family":"Garcia","sequence":"additional","affiliation":[{"name":"University of Konstanz, Universitaetstrasse 10, Konstanz, 78457, Germany."},{"name":"Complexity Science Hub, Josefstaedter Strasse 39, Vienna, 1080, Austria."}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"221","reference":[{"key":"e_1_3_2_2_2","doi-asserted-by":"crossref","unstructured":"J. Henrich \u201cThe secret of our success: How culture is driving human evolution domesticating our species and making us smarter \u201d in The Secret of our Success (Princeton Univ. Press 2015).","DOI":"10.2307\/j.ctvc77f0d"},{"key":"e_1_3_2_3_2","unstructured":"R. I. M. Dunbar Grooming Gossip and the Evolution of Language (Harvard Univ. Press 1996)."},{"key":"e_1_3_2_4_2","unstructured":"M. Chen J. Tworek H. Jun Q. Yuan Henrique Ponde de Oliveira Pinto J. Kaplan H. Edwards Y. Burda N. Joseph G. Brockman A. Ray R. Puri G. Krueger M. Petrov H. Khlaaf G. Sastry P. Mishkin B. Chan S. Gray N. Ryder M. Pavlov A. Power L. Kaiser M. Bavarian C. Winter P. Tillet F. P. Such D. Cummings M. Plappert F. Chantzis E. Barnes A. Herbert-Voss W. H. Guss A. Nichol A. Paino N. Tezak J. Tang I. Babuschkin S. Balaji S. Jain W. Saunders C. Hesse A. N. Carr J. Leike J. Achiam V. Misra E. Morikawa A. Radford M. Knight M. Brundage M. Murati K. Mayer P. Welinder B. M. Grew D. Amodei S. M. Candlish I. Sutskever W. Zaremba Evaluating large language models trained on code. arXiv:2107.03374 [cs.LG] (2021)."},{"key":"e_1_3_2_5_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-023-06291-2"},{"key":"e_1_3_2_6_2","doi-asserted-by":"publisher","DOI":"10.1098\/rsta.2023.0254"},{"key":"e_1_3_2_7_2","doi-asserted-by":"publisher","DOI":"10.1140\/epjds\/s13688-023-00427-0"},{"key":"e_1_3_2_8_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-023-06792-0"},{"key":"e_1_3_2_9_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-023-06924-6"},{"key":"e_1_3_2_10_2","unstructured":"J. Achiam S. Adler S. Agarwal L. Ahmad I. Akkaya F. L. Aleman D. Almeida J. Altenschmidt S. Altman S. Anadkat R. Avila I. Babuschkin S. Balaji V. Balcom P. Baltescu H. Bao M. Bavarian J. Belgum I. Bello J. Berdine G. Bernadett-Shapiro C. Berner L. Bogdonoff O. Boiko M. Boyd A.-L. Brakman G. Brockman T. Brooks M. Brundage K. Button T. Cai R. Campbell A. Cann B. Carey C. Carlson R. Carmichael B.Chan C. Chang F. Chantzis D. Chen S. Chen R. Chen J. Chen M. Chen B. Chess C. Cho C. Chu H. W. Chung D. Cummings J. Currier Y. Dai C. Decareaux T. Degry N. Deutsch D. Deville A. Dhar D. Dohan S. Dowling S. Dunning A. Ecoffet A. Eleti T. Eloundou D. Farhi L. Fedus N. Felix S. P. Fishman J. Forte I. Fulford L. Gao E. Georges C. Gibson V. Goel T. Gogineni G. Goh R. Gontijo-Lopes J. Gordon M. Grafstein S. Gray R. Greene J. Gross S. S. Gu Y. Guo C. Hallacy J. Han J. Harris Y. He M. Heaton J. Heidecke C. Hesse A. Hickey W. Hickey P. Hoeschele B. Houghton K. Hsu S. Hu X. Hu J. Huizinga S. Jain S. Jain J. Jang A. Jiang R. Jiang H. Jin D. Jin S. Jomoto B. Jonn H. Jun T. Kaftan \u0141. Kaiser A. Kamali I. Kanitscheider N. S. Keskar T. Khan L. Kilpatrick J. W. Kim C. Kim Y. Kim J. H. Kirchner J. Kiros M. Knight D. Kokotajlo \u0141. Kondraciuk A. Kondrich A. Konstantinidis K. Kosic G. Krueger V. Kuo M. Lampe I. Lan T. Lee J. Leike J. Leung D. Levy C. M. Li R. Lim M. Lin S. Lin M. Litwin T. Lopez R. Lowe P. Lue A. Makanju K. Malfacini S. Manning T. Markov Y. Markovski B. Martin K. Mayer A. Mayne B. M. Grew S. M. McKinney C. M. Leavey P. M. Millan J. M. Neil D. Medina A. Mehta J. Menick L. Metz A. Mishchenko P. Mishkin V. Monaco E. Morikawa D. Mossing T. Mu M. Murati O. Murk D. M\u00e9ly A. Nair R. Nakano R. Nayak A. Neelakantan R. Ngo H. Noh L. Ouyang C. O\u2019Keefe J. Pachocki A. Paino J. Palermo A. Pantuliano G. Parascandolo J. Parish E. Parparita A. Passos M. Pavlov A. Peng A. Perelman Filipe de Avila Belbute Peres M. Petrov H. P. de Oliveira Pinto M. Pokorny M. Pokrass V. H. Pong T. Powell A. Power B. Power E. Proehl R. Puri A. Radford J. Rae A. Ramesh C. Raymond F. Real K. Rimbach C. Ross B. Rotsted H. Roussez N. Ryder M. Saltarelli T. Sanders S. Santurkar G. Sastry H. Schmidt D. Schnurr J. Schulman D. Selsam K. Sheppard T. Sherbakov J. Shieh S. Shoker P. Shyam S. Sidor E. Sigler M. Simens J. Sitkin K. Slama I. Sohl B. Sokolowsky Y. Song N. Staudacher F. P. Such N. Summers I. Sutskever J. Tang N. Tezak M. B. Thompson P. Tillet A. Tootoonchian E. Tseng P. Tuggle N. Turley J. Tworek J. F. C. Uribe A. Vallone A. Vijayvergiya C. Voss C. Wainwright J. J. Wang A. Wang B. Wang J. Ward J. Wei C. J. Weinmann A. Welihinda P. Welinder J. Weng L. Weng M. Wiethoff D. Willner C. Winter S. Wolrich H. Wong L. Workman S. Wu J. Wu M. Wu K. Xiao T. Xu S. Yoo K. Yu Q. Yuan W. Zaremba R. Zellers C. Zhang M. Zhang S. Zhao T. Zheng J. Zhuang W. Zhuk B. Zoph Gpt-4 technical report. arXiv:2303.08774 [cs.CL] (2023)."},{"key":"e_1_3_2_11_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11432-024-4222-0"},{"key":"e_1_3_2_12_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.1193147"},{"key":"e_1_3_2_13_2","unstructured":"Q. Wu G. Bansal J. Zhang Y. Wu B. Li E. Zhu L. Jiang X. Zhang S. Zhang J. Liu A. H. Awadallah R. W. White D. Burger C. Wang \u201cAutogen: Enabling next-gen LLM applications via multi-agent conversations \u201d in First Conference on Language Modeling (Conference on Language Modeling; 2024)."},{"key":"e_1_3_2_14_2","unstructured":"A. Q. Jiang A. Sablayrolles A. Roux A. Mensch B. Savary C. Bamford D. S. Chaplot Diego de las Casas E. B. Hanna F. Bressand G. Lengyel G. Bour G. Lample L. R. Lavaud L. Saulnier M.-A. Lachaux P. Stock S. Subramanian S. Yang S. Antoniak T. L. Scao T. Gervet T. Lavril T. Wang T. Lacroix W. E. Sayed Mixtral of experts. arXiv:2401.04088 [cs.LG] (2024)."},{"key":"e_1_3_2_15_2","unstructured":"X. Guo K. Huang J. Liu W. Fan N. V\u00e9lez Q. Wu H. Wang T. L. Griffiths M. Wang Embodied llm agents learn to cooperate in organized teams. arXiv:2403.12482 [cs.AI] (2024)."},{"key":"e_1_3_2_16_2","unstructured":"Z. Liu Y. Zhang P. Li Y. Liu D. Yang Dynamic llm-agent network: An llm-agent collaboration framework with agent team optimization. arXiv:2310.02170 [cs.CL] (2023)."},{"key":"e_1_3_2_17_2","article-title":"Hugginggpt: Solving ai tasks with chatgpt and its friends in hugging face","volume":"36","author":"Shen Y.","year":"2024","unstructured":"Y. Shen, K. Song, X. Tan, D. Li, W. Lu, Y. Zhuang, Hugginggpt: Solving ai tasks with chatgpt and its friends in hugging face. Adv. Neural Inf. Proces. Syst. 36, 38154\u201338180 (2024).","journal-title":"Adv. Neural Inf. Proces. Syst."},{"key":"e_1_3_2_18_2","doi-asserted-by":"publisher","DOI":"10.1038\/srep02627"},{"key":"e_1_3_2_19_2","unstructured":"A. S. Vezhnevets J. P. Agapiou A. Aharon R. Ziv J. Matyas E. A. Du\u00e9\u00f1ez-Guzm\u00e1n W. A. Cunningham S. Osindero D. Karmon J. Z. Leibo Generative agent-based modeling with actions grounded in physical social or digital space using Concordia. arXiv:2312.03664 [cs.AI] (2023)."},{"key":"e_1_3_2_20_2","doi-asserted-by":"publisher","DOI":"10.1038\/nature03236"},{"key":"e_1_3_2_21_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-019-1138-y"},{"key":"e_1_3_2_22_2","unstructured":"G. V. Aher R. I. Arriaga A. T. Kalai \u201cUsing large language models to simulate multiple humans and replicate human subject studies \u201d in Proceedings of the 40th International Conference on Machine Learning (PMLR; 2023) vol. 202 pp. 337\u2013371."},{"key":"e_1_3_2_23_2","doi-asserted-by":"publisher","DOI":"10.1017\/pan.2023.2"},{"key":"e_1_3_2_24_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2309583120"},{"key":"e_1_3_2_25_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2218523120"},{"key":"e_1_3_2_26_2","doi-asserted-by":"publisher","DOI":"10.1177\/17456916231214460"},{"key":"e_1_3_2_27_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41562-024-01882-z"},{"key":"e_1_3_2_28_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.adi1778"},{"key":"e_1_3_2_29_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2314021121"},{"key":"e_1_3_2_30_2","doi-asserted-by":"crossref","unstructured":"Y. Lu A. Aleta C. Du L. Shi Y. Moreno Generative agent-based models for complex systems research: A review. arXiv:2408.09175 [physics.soc-ph] (2024).","DOI":"10.1016\/j.plrev.2024.10.013"},{"key":"e_1_3_2_31_2","unstructured":"S. Lai Y. Potter J. Kim R. Zhuang D. Song J. Evans \u201cPosition: Evolving AI collectives enhance human diversity and enable self-regulation \u201d in Proceedings of the 41st International Conference on Machine Learning (PMLR; 2024)."},{"key":"e_1_3_2_32_2","doi-asserted-by":"crossref","unstructured":"M. Papachristou Y. Yuan Network formation and dynamics among multi-LLMs. arXiv:2402.10659 [cs.SI] (2024).","DOI":"10.2139\/ssrn.5410522"},{"key":"e_1_3_2_33_2","doi-asserted-by":"crossref","unstructured":"S. Chang A. Chaszczewicz E. Wang M. Josifovska E. Pierson J. Leskovec LLMs generate structurally realistic social networks but overestimate political homophily. arXiv:2408.16629 [cs.CY] (2024).","DOI":"10.1609\/icwsm.v19i1.35820"},{"key":"e_1_3_2_34_2","doi-asserted-by":"crossref","unstructured":"A. C. Kozlowski H. Kwon J. A. Evans In silico sociology: Forecasting COVID-19 polarization with large language models. arXiv:2407.11190 [cs.CY] (2024).","DOI":"10.31235\/osf.io\/7dfbc"},{"key":"e_1_3_2_35_2","unstructured":"G. De Marzo L. Pietronero D. Garcia Emergence of scale-free networks in social interactions among large language models. arXiv:2312.06619 [physics.soc-ph] (2023)."},{"key":"e_1_3_2_36_2","unstructured":"P. T\u00f6rnberg D. Valeeva J. Uitermark C. Bail Simulating social media using large language models to evaluate alternative news feed algorithms. arXiv:2310.05984 [cs.SI] (2023)."},{"key":"e_1_3_2_37_2","unstructured":"J. Piao Z. Lu C. Gao F. Xu Q. Hu F. P. Santos Y. Li J. Evans Emergence of human-like polarization among large language model agents. arXiv:2501.05171 [cs.SI] (2025)."},{"key":"e_1_3_2_38_2","doi-asserted-by":"crossref","unstructured":"J. S. Park J. O\u2019Brien C. J. Cai M. R. Morris P. Liang M. S. Bernstein \u201cGenerative agents: Interactive simulacra of human behavior \u201d in Proceedings of the 36th Annual ACM Symposium on User Interface Software And Technology (Association for Computing Machinery New York NY; 2023) pp. 1\u201322.","DOI":"10.1145\/3586183.3606763"},{"key":"e_1_3_2_39_2","doi-asserted-by":"crossref","unstructured":"Y.-S. Chuang A. Goyal N. Harlalka S. Suresh R. Hawkins S. Yang D. Shah J. Hu T. T. Rogers Simulating opinion dynamics with networks of llm-based agents. arXiv:2311.09618 [physics.soc-ph] (2023).","DOI":"10.18653\/v1\/2024.findings-naacl.211"},{"key":"e_1_3_2_40_2","unstructured":"E. Cau V. Pansanella D. Pedreschi G. Rossetti Language-driven opinion dynamics in agent-based simulations with LLMs. arXiv:2502.19098 [cs.SI] (2025)."},{"key":"e_1_3_2_41_2","doi-asserted-by":"publisher","DOI":"10.1126\/sciadv.adu9368"},{"key":"e_1_3_2_42_2","unstructured":"C. Barrie P. T\u00f6rnberg Emergent LLM behaviors are observationally equivalent to data leakage. arXiv:2505.23796 [cs.CL] (2025)."},{"key":"e_1_3_2_43_2","doi-asserted-by":"publisher","DOI":"10.1038\/nature13998"},{"key":"e_1_3_2_44_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1621067114"},{"key":"e_1_3_2_45_2","doi-asserted-by":"publisher","DOI":"10.1103\/RevModPhys.81.591"},{"key":"e_1_3_2_46_2","doi-asserted-by":"publisher","DOI":"10.1088\/0143-0807\/34\/6\/1555"},{"key":"e_1_3_2_47_2","doi-asserted-by":"publisher","DOI":"10.1017\/S0140525X05000129"},{"key":"e_1_3_2_48_2","doi-asserted-by":"publisher","DOI":"10.1002\/mde.2555"},{"key":"e_1_3_2_49_2","doi-asserted-by":"publisher","DOI":"10.1063\/1.1703954"},{"key":"e_1_3_2_50_2","doi-asserted-by":"publisher","DOI":"10.1016\/0047-2484(92)90081-J"},{"key":"e_1_3_2_51_2","unstructured":"D. Hendrycks C. Burns S. Basart A. Zou M. Mazeika D. Song J. Steinhardt Measuring massive multitask language understanding. arXiv:2009.03300 [cs.CY] (2020)."},{"key":"e_1_3_2_52_2","doi-asserted-by":"crossref","unstructured":"S. Zheng Y. Zhang Y. Zhu C. Xi P. Gao X. Zhou Kevin Chen-Chuan Chang Gpt-fathom: Benchmarking large language models to decipher the evolutionary path towards GPT-4 and beyond. arXiv:2309.16583 [cs.CL] (2023).","DOI":"10.18653\/v1\/2024.findings-naacl.87"},{"key":"e_1_3_2_53_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41559-017-0112"},{"key":"e_1_3_2_54_2","doi-asserted-by":"publisher","DOI":"10.1098\/rsbl.2021.0158"},{"key":"e_1_3_2_55_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1713496115"},{"key":"e_1_3_2_56_2","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0022656"},{"key":"e_1_3_2_57_2","unstructured":"G. De Marzo D. Garcia Collective behavior of AI agents: The case of moltbook. arXiv:2602.09270 [physics.soc-ph] (2026)."},{"key":"e_1_3_2_58_2","doi-asserted-by":"crossref","unstructured":"F. Fadaei J. C. Moran T. Yasseri Gender dynamics and homophily in a social network of LLM agents. arXiv:2602.02606 [cs.SI] (2026).","DOI":"10.31235\/osf.io\/5jrfn_v1"},{"key":"e_1_3_2_59_2","doi-asserted-by":"publisher","DOI":"10.1086\/261849"}],"container-title":["Science Advances"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.science.org\/doi\/pdf\/10.1126\/sciadv.aea6091","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T18:01:01Z","timestamp":1786730461000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.science.org\/doi\/10.1126\/sciadv.aea6091"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,14]]},"references-count":58,"journal-issue":{"issue":"33","published-print":{"date-parts":[[2026,8,14]]}},"alternative-id":["10.1126\/sciadv.aea6091"],"URL":"https:\/\/doi.org\/10.1126\/sciadv.aea6091","relation":{},"ISSN":["2375-2548"],"issn-type":[{"value":"2375-2548","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,14]]},"assertion":[{"value":"2025-07-15","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-07-07","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-08-14","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"eaea6091"}}