{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,24]],"date-time":"2026-08-24T09:31:08Z","timestamp":1787563868460,"version":"build-2736575974"},"reference-count":41,"publisher":"Cambridge University Press (CUP)","issue":"3","license":[{"start":{"date-parts":[[2023,2,21]],"date-time":"2023-02-21T00:00:00Z","timestamp":1676937600000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/www.cambridge.org\/core\/terms"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Polit. Anal."],"published-print":{"date-parts":[[2023,7]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    We propose and explore the possibility that language models can be studied as effective proxies for specific human subpopulations in social science research. Practical and research applications of artificial intelligence tools have sometimes been limited by problematic biases (such as racism or sexism), which are often treated as uniform properties of the models. We show that the \u201calgorithmic bias\u201d within one such tool\u2014the GPT-3 language model\u2014is instead both fine-grained and demographically correlated, meaning that proper conditioning will cause it to accurately emulate response distributions from a wide variety of human subgroups. We term this property\n                    <jats:italic>algorithmic fidelity<\/jats:italic>\n                    and explore its extent in GPT-3. We create \u201csilicon samples\u201d by conditioning the model on thousands of sociodemographic backstories from real human participants in multiple large surveys conducted in the United States. We then compare the silicon and human samples to demonstrate that the information contained in GPT-3 goes far beyond surface similarity. It is nuanced, multifaceted, and reflects the complex interplay between ideas, attitudes, and sociocultural context that characterize human attitudes. We suggest that language models with sufficient algorithmic fidelity thus constitute a novel and powerful tool to advance understanding of humans and society across a variety of disciplines.\n                  <\/jats:p>","DOI":"10.1017\/pan.2023.2","type":"journal-article","created":{"date-parts":[[2023,2,21]],"date-time":"2023-02-21T05:53:18Z","timestamp":1676958798000},"page":"337-351","source":"Crossref","is-referenced-by-count":631,"title":["Out of One, Many: Using Language Models to Simulate Human Samples"],"prefix":"10.1017","volume":"31","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-3109-2537","authenticated-orcid":false,"given":"Lisa P.","family":"Argyle","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ethan C.","family":"Busby","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nancy","family":"Fulda","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1635-8210","authenticated-orcid":false,"given":"Joshua R.","family":"Gubler","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Christopher","family":"Rytting","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"David","family":"Wingate","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"56","published-online":{"date-parts":[[2023,2,21]]},"reference":[{"key":"S1047198723000025_r24","doi-asserted-by":"publisher","DOI":"10.1093\/poq\/nfs038"},{"key":"S1047198723000025_r16","doi-asserted-by":"publisher","DOI":"10.1146\/annurev-polisci-060418-042842"},{"key":"S1047198723000025_r8","volume-title":"Voting: A Study of Opinion Formation in a Presidential Campaign","author":"Berelson","year":"1954"},{"key":"S1047198723000025_r15","volume-title":"Mathematical Methods of Statistics","author":"Cram\u00e9r","year":"1946"},{"key":"S1047198723000025_r35","doi-asserted-by":"publisher","DOI":"10.1017\/pan.2019.26"},{"key":"S1047198723000025_r31","first-page":"2218","article-title":"Bias In, Bias Out","volume":"128","author":"Mayson","year":"2018","journal-title":"Yale Law Journal"},{"key":"S1047198723000025_r9","doi-asserted-by":"publisher","DOI":"10.1017\/S0003055404001315"},{"key":"S1047198723000025_r18","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P19-1285"},{"key":"S1047198723000025_r19","doi-asserted-by":"publisher","DOI":"10.1146\/annurev-polisci-020614-095051"},{"key":"S1047198723000025_r25","doi-asserted-by":"publisher","DOI":"10.1017\/9781108645157"},{"key":"S1047198723000025_r33","first-page":"9","article-title":"Language Models Are Unsupervised Multitask Learners","volume":"1","author":"Radford","year":"2019","journal-title":"OpenAI Blog"},{"key":"S1047198723000025_r39","volume-title":"Bit by Bit: Social Research in the Digital Age","author":"Salganik","year":"2017"},{"key":"S1047198723000025_r6","doi-asserted-by":"crossref","unstructured":"Bender, E. M. , Gebru, T. , McMillan-Major, A. , and Shmitchell, S. . 2021. \u201cOn the Dangers of Stochastic Parrots: Can Language Models Be Too Big?\u201d In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, 610\u2013623.","DOI":"10.1145\/3442188.3445922"},{"key":"S1047198723000025_r36","doi-asserted-by":"publisher","DOI":"10.1086\/715162"},{"key":"S1047198723000025_r40","doi-asserted-by":"crossref","first-page":"238","DOI":"10.1111\/j.2517-6161.1951.tb00088.x","article-title":"The Interpretation of Interaction in Contingency Tables","volume":"13","author":"Simpson","year":"1951","journal-title":"Journal of the Royal Statistical Society, Series B"},{"key":"S1047198723000025_r5","first-page":"671","article-title":"Big Data\u2019s Disparate Impact","volume":"104","author":"Barocas","year":"2016","journal-title":"California Law Review"},{"key":"S1047198723000025_r23","doi-asserted-by":"publisher","DOI":"10.1146\/annurev.polisci.7.012003.104859"},{"key":"S1047198723000025_r28","volume-title":"Facing the Challenge of Democracy: Explorations in the Analysis of Public Opinion and Political Participation","author":"Magleby","year":"2011"},{"key":"S1047198723000025_r34","first-page":"1","article-title":"Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer","volume":"21","author":"Raffel","year":"2020","journal-title":"Journal of Machine Learning Research"},{"key":"S1047198723000025_r10","unstructured":"Brown, T. B. , et al. 2020. \u201cLanguage Models Are Few-Shot Learners.\u201d Advances in Neural Information Processing Systems 33: 1\u201325."},{"key":"S1047198723000025_r4","doi-asserted-by":"publisher","DOI":"10.1017\/pan.2020.8"},{"key":"S1047198723000025_r21","doi-asserted-by":"publisher","DOI":"10.1017\/pan.2018.11"},{"key":"S1047198723000025_r11","doi-asserted-by":"publisher","DOI":"10.1146\/annurev.polisci.12.040507.142213"},{"key":"S1047198723000025_r30","doi-asserted-by":"publisher","DOI":"10.7208\/chicago\/9780226524689.001.0001"},{"key":"S1047198723000025_r7","doi-asserted-by":"publisher","DOI":"10.1111\/ajps.12423"},{"key":"S1047198723000025_r38","doi-asserted-by":"publisher","DOI":"10.1007\/s11109-018-9457-5"},{"key":"S1047198723000025_r2","unstructured":"ANES. 2021. \u201cAmerican National Election Studies.\u201d https:\/\/electionstudies.org\/about-us\/."},{"key":"S1047198723000025_r29","unstructured":"Marcus, G. 2020. \u201cThe Next Decade in AI: Four Steps towards Robust Artificial Intelligence.\u201d Preprint, arXiv:2002.06177."},{"key":"S1047198723000025_r37","volume-title":"Guide for Conducting Risk Assessments (Nist Sp-800-30rev1)","author":"Ross","year":"2012"},{"key":"S1047198723000025_r41","volume-title":"From Protest to Politics: The New Black Voters in American Elections","author":"Tate","year":"1994"},{"key":"S1047198723000025_r27","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9781316471050"},{"key":"S1047198723000025_r3","doi-asserted-by":"publisher","DOI":"10.7910\/DVN\/JPV20K"},{"key":"S1047198723000025_r20","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1720347115"},{"key":"S1047198723000025_r26","doi-asserted-by":"crossref","DOI":"10.1525\/9780520912236","volume-title":"The Myth of the Independent Voter","author":"Keith","year":"1992"},{"key":"S1047198723000025_r12","doi-asserted-by":"publisher","DOI":"10.1126\/science.aal4230"},{"key":"S1047198723000025_r17","doi-asserted-by":"publisher","DOI":"10.7208\/chicago\/9780226349251.001.0001"},{"key":"S1047198723000025_r14","doi-asserted-by":"publisher","DOI":"10.1177\/2053168018822174"},{"key":"S1047198723000025_r22","doi-asserted-by":"publisher","DOI":"10.1146\/annurev-polisci-053119-015921"},{"key":"S1047198723000025_r1","unstructured":"Adiwardana, D. , et al. 2020. \u201cTowards a Human-Like Open-Domain Chatbot.\u201d Preprint, arXiv:2001.09977."},{"key":"S1047198723000025_r32","doi-asserted-by":"publisher","DOI":"10.7189\/jogh.09.020318"},{"key":"S1047198723000025_r13","volume-title":"The American Voter","author":"Campbell","year":"1960"}],"container-title":["Political Analysis"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.cambridge.org\/core\/services\/aop-cambridge-core\/content\/view\/S1047198723000025","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,10,14]],"date-time":"2024-10-14T15:02:24Z","timestamp":1728918144000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.cambridge.org\/core\/product\/identifier\/S1047198723000025\/type\/journal_article"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,2,21]]},"references-count":41,"journal-issue":{"issue":"3","published-print":{"date-parts":[[2023,7]]}},"alternative-id":["S1047198723000025"],"URL":"https:\/\/doi.org\/10.1017\/pan.2023.2","relation":{},"ISSN":["1047-1987","1476-4989"],"issn-type":[{"value":"1047-1987","type":"print"},{"value":"1476-4989","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,2,21]]}}}