{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T15:37:37Z","timestamp":1787067457144,"version":"build-2736575974"},"reference-count":79,"publisher":"National Academy of Sciences","issue":"51","license":[{"start":{"date-parts":[[2023,12,12]],"date-time":"2023-12-12T00:00:00Z","timestamp":1702339200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100010225","name":"MOST | NSFC | National Outstanding Youth Foundation of China","doi-asserted-by":"publisher","award":["72222005"],"award-info":[{"award-number":["72222005"]}],"id":[{"id":"10.13039\/501100010225","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100004147","name":"Tsinghua University","doi-asserted-by":"publisher","award":["2022Z04W01032"],"award-info":[{"award-number":["2022Z04W01032"]}],"id":[{"id":"10.13039\/501100004147","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100005950","name":"Hong Kong University of Science and Technology","doi-asserted-by":"publisher","award":["00000"],"award-info":[{"award-number":["00000"]}],"id":[{"id":"10.13039\/501100005950","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.pnas.org"],"crossmark-restriction":true},"short-container-title":["Proc. Natl. Acad. Sci. U.S.A."],"published-print":{"date-parts":[[2023,12,19]]},"abstract":"<jats:p>As large language models (LLMs) like GPT become increasingly prevalent, it is essential that we assess their capabilities beyond language processing. This paper examines the economic rationality of GPT by instructing it to make budgetary decisions in four domains: risk, time, social, and food preferences. We measure economic rationality by assessing the consistency of GPT\u2019s decisions with utility maximization in classic revealed preference theory. We find that GPT\u2019s decisions are largely rational in each domain and demonstrate higher rationality score than those of human subjects in a parallel experiment and in the literature. Moreover, the estimated preference parameters of GPT are slightly different from human subjects and exhibit a lower degree of heterogeneity. We also find that the rationality scores are robust to the degree of randomness and demographic settings such as age and gender but are sensitive to contexts based on the language frames of the choice situations. These results suggest the potential of LLMs to make good decisions and the need to further understand their capabilities, limitations, and underlying mechanisms.<\/jats:p>","DOI":"10.1073\/pnas.2316205120","type":"journal-article","created":{"date-parts":[[2023,12,12]],"date-time":"2023-12-12T13:38:54Z","timestamp":1702388334000},"update-policy":"https:\/\/doi.org\/10.1073\/pnas.cm10313","source":"Crossref","is-referenced-by-count":139,"title":["The emergence of economic rationality of GPT"],"prefix":"10.1073","volume":"120","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-7473-0263","authenticated-orcid":false,"given":"Yiting","family":"Chen","sequence":"first","affiliation":[{"name":"Department of Economics, Lingnan University, Hong Kong, China HKG"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tracy Xiao","family":"Liu","sequence":"additional","affiliation":[{"name":"Department of Economics, School of Economics and Management, National Center for Economic Research at Tsinghua University, Tsinghua University, Beijing 100084, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8899-5393","authenticated-orcid":false,"given":"You","family":"Shan","sequence":"additional","affiliation":[{"name":"Department of Economics, School of Economics and Management, Tsinghua University, Beijing 100084, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7661-8948","authenticated-orcid":false,"given":"Songfa","family":"Zhong","sequence":"additional","affiliation":[{"name":"Department of Economics, Hong Kong University of Science and Technology, Hong Kong, China HKG"},{"name":"Department of Economics, National University of Singapore, Singapore 117570, Singapore"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"341","published-online":{"date-parts":[[2023,12,12]]},"reference":[{"key":"e_1_3_4_1_2","first-page":"1","article-title":"Attention is all you need","volume":"30","author":"Vaswani A.","year":"2017","unstructured":"A. Vaswani , Attention is all you need. Adv. Neural Inf. Process. Syst. 30, 1\u201311 (2017).","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"e_1_3_4_2_2","first-page":"1877","article-title":"Language models are few-shot learners","volume":"33","author":"Brown T.","year":"2020","unstructured":"T. Brown , Language models are few-shot learners. Adv. Neural Inf. Process. Syst. 33, 1877\u20131901 (2020).","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"e_1_3_4_3_2","unstructured":"M. Chen . Evaluating large language models trained on code. arXiv [Preprint] (2021). http:\/\/arxiv.org\/abs\/2107.03374 (Accessed 25 November 2023)."},{"key":"e_1_3_4_4_2","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v34i09.7098"},{"key":"e_1_3_4_5_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2123433119"},{"key":"e_1_3_4_6_2","unstructured":"M. Kosinski Theory of mind may have spontaneously emerged in large language models. arXiv [Preprint] (2023). http:\/\/arxiv.org\/abs\/2302.02083 (Accessed 25 November 2023)."},{"key":"e_1_3_4_7_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2218523120"},{"key":"e_1_3_4_8_2","unstructured":"P. S. Park P. Schoenegger C. Zhu Artificial intelligence in psychology research. arXiv [Preprint] (2023). http:\/\/arxiv.org\/abs\/2302.07267 (Accessed 25 November 2023)."},{"key":"e_1_3_4_9_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41562-023-01659-w"},{"key":"e_1_3_4_10_2","doi-asserted-by":"crossref","unstructured":"J. J. Horton Large language models as simulated economic agents: What can we learn from homo silicus? NBER Working Paper (2023).","DOI":"10.3386\/w31122"},{"key":"e_1_3_4_11_2","doi-asserted-by":"publisher","DOI":"10.2307\/2525382"},{"key":"e_1_3_4_12_2","doi-asserted-by":"publisher","DOI":"10.2307\/2525845"},{"key":"e_1_3_4_13_2","doi-asserted-by":"publisher","DOI":"10.2307\/2548836"},{"key":"e_1_3_4_14_2","doi-asserted-by":"publisher","DOI":"10.2307\/1912771"},{"key":"e_1_3_4_15_2","doi-asserted-by":"publisher","DOI":"10.1016\/0304-4076(90)90051-T"},{"key":"e_1_3_4_16_2","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9781316104293"},{"key":"e_1_3_4_17_2","doi-asserted-by":"publisher","DOI":"10.1257\/aer.20150947"},{"key":"e_1_3_4_18_2","doi-asserted-by":"publisher","DOI":"10.1111\/1468-0262.00302"},{"key":"e_1_3_4_19_2","doi-asserted-by":"publisher","DOI":"10.1257\/aer.102.7.3333"},{"key":"e_1_3_4_20_2","doi-asserted-by":"publisher","DOI":"10.3982\/QE243"},{"key":"e_1_3_4_21_2","doi-asserted-by":"publisher","DOI":"10.1257\/aer.97.5.1921"},{"key":"e_1_3_4_22_2","doi-asserted-by":"publisher","DOI":"10.1257\/aer.104.6.1518"},{"key":"e_1_3_4_23_2","doi-asserted-by":"publisher","DOI":"10.1257\/aer.91.5.1539"},{"key":"e_1_3_4_24_2","doi-asserted-by":"publisher","DOI":"10.1086\/697741"},{"key":"e_1_3_4_25_2","doi-asserted-by":"publisher","DOI":"10.1257\/aer.97.5.1858"},{"key":"e_1_3_4_26_2","doi-asserted-by":"publisher","DOI":"10.1111\/1468-0262.00394"},{"key":"e_1_3_4_27_2","doi-asserted-by":"publisher","DOI":"10.3982\/ECTA6069"},{"key":"e_1_3_4_28_2","doi-asserted-by":"publisher","DOI":"10.1086\/665011"},{"key":"e_1_3_4_29_2","doi-asserted-by":"publisher","DOI":"10.1162\/REST_a_00542"},{"key":"e_1_3_4_30_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.geb.2019.05.003"},{"key":"e_1_3_4_31_2","doi-asserted-by":"publisher","DOI":"10.1086\/503550"},{"key":"e_1_3_4_32_2","doi-asserted-by":"publisher","DOI":"10.1111\/j.1465-7295.1975.tb01101.x"},{"key":"e_1_3_4_33_2","doi-asserted-by":"publisher","DOI":"10.1162\/rest_a_00785"},{"key":"e_1_3_4_34_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.geb.2023.10.005"},{"key":"e_1_3_4_35_2","doi-asserted-by":"publisher","DOI":"10.1257\/aer.20140481"},{"key":"e_1_3_4_36_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.jpubeco.2017.08.010"},{"key":"e_1_3_4_37_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.aar6987"},{"key":"e_1_3_4_38_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1705451114"},{"key":"e_1_3_4_39_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2112726119"},{"key":"e_1_3_4_40_2","doi-asserted-by":"publisher","DOI":"10.1257\/000282803322655392"},{"key":"e_1_3_4_41_2","doi-asserted-by":"publisher","DOI":"10.1257\/aer.91.3.351"},{"key":"e_1_3_4_42_2","unstructured":"S. Corbett-Davies S. Goel The measure and mismeasure of fairness: A critical review of fair machine learning. arXiv [Preprint] (2018). http:\/\/arxiv.org\/abs\/1808.00023 (Accessed 25 November 2023)."},{"key":"e_1_3_4_43_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-019-1138-y"},{"key":"e_1_3_4_44_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2215907120"},{"key":"e_1_3_4_45_2","doi-asserted-by":"crossref","unstructured":"A. Borji A categorical archive of chatGPT failures. arXiv [Preprint] (2023). http:\/\/arxiv.org\/abs\/2302.03494 (Accessed 25 November 2023).","DOI":"10.21203\/rs.3.rs-2895792\/v1"},{"key":"e_1_3_4_46_2","doi-asserted-by":"crossref","unstructured":"Y. Chen M. Andiappan T. Jenkin A. Ovchinnikov A manager and an AI walk into a bar: Does chatGPT make biased decisions like we do? Available at SSRN 4380365 (2023).","DOI":"10.2139\/ssrn.4380365"},{"key":"e_1_3_4_47_2","unstructured":"K. Mahowald . Dissociating language and thought in large language models: A cognitive perspective. arXiv [Preprint] (2023). http:\/\/arxiv.org\/abs\/2301.06627 (Accessed 25 November 2023)."},{"key":"e_1_3_4_48_2","first-page":"11785","article-title":"Capturing failures of large language models via human cognitive biases","volume":"35","author":"Jones E.","year":"2022","unstructured":"E. Jones, J. Steinhardt, Capturing failures of large language models via human cognitive biases. Adv. Neural Inf. Process. Syst. 35, 11785\u201311799 (2022).","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"e_1_3_4_49_2","doi-asserted-by":"crossref","unstructured":"J. Brand A. Israeli D. Ngwe Using GPT for market research. Available at SSRN 4395751 (2023).","DOI":"10.2139\/ssrn.4395751"},{"key":"e_1_3_4_50_2","volume-title":"Deep Learning","author":"Goodfellow I.","year":"2016","unstructured":"I. Goodfellow, Y. Bengio, A. Courville, Deep Learning (MIT Press, 2016)."},{"key":"e_1_3_4_51_2","doi-asserted-by":"crossref","unstructured":"M. Bommarito II D. M. Katz GPT takes the bar exam. arXiv [Preprint] (2022). http:\/\/arxiv.org\/abs\/2212.14402 (Accessed 25 November 2023).","DOI":"10.2139\/ssrn.4314839"},{"key":"e_1_3_4_52_2","doi-asserted-by":"publisher","DOI":"10.1093\/ej\/ueaa052"},{"key":"e_1_3_4_53_2","unstructured":"M. Chen T. X. Liu Y. Shan S. Zhong Y. Zhou The consistency of rationality measures. Working Paper (2023)."},{"key":"e_1_3_4_54_2","doi-asserted-by":"publisher","DOI":"10.2307\/2296210"},{"key":"e_1_3_4_55_2","doi-asserted-by":"publisher","DOI":"10.1257\/mic.5.1.28"},{"key":"e_1_3_4_56_2","first-page":"89","article-title":"Determining all maximal data subsets consistent with revealed preference","volume":"19","author":"Houtman M.","year":"1985","unstructured":"M. Houtman, J. Maks, Determining all maximal data subsets consistent with revealed preference. Kwantitatieve Methoden 19, 89\u2013104 (1985).","journal-title":"Kwantitatieve Methoden"},{"key":"e_1_3_4_57_2","doi-asserted-by":"publisher","DOI":"10.2307\/2938223"},{"key":"e_1_3_4_58_2","doi-asserted-by":"publisher","DOI":"10.2307\/1913608"},{"key":"e_1_3_4_59_2","doi-asserted-by":"publisher","DOI":"10.1257\/aer.101.6.2782"},{"key":"e_1_3_4_60_2","doi-asserted-by":"publisher","DOI":"10.1093\/jeea\/jvad028"},{"key":"e_1_3_4_61_2","doi-asserted-by":"publisher","DOI":"10.1257\/aer.101.2.664"},{"key":"e_1_3_4_62_2","first-page":"237","article-title":"Human decisions and machine predictions","volume":"133","author":"Kleinberg J.","year":"2018","unstructured":"J. Kleinberg, H. Lakkaraju, J. Leskovec, J. Ludwig, S. Mullainathan, Human decisions and machine predictions. Quart. J. Econ. 133, 237\u2013293 (2018).","journal-title":"Quart. J. Econ."},{"key":"e_1_3_4_63_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-022-01804-4"},{"key":"e_1_3_4_64_2","unstructured":"Y. Kawaguchi . Description and discussion on DCASE 2021 challenge task 2: Unsupervised anomalous sound detection for machine condition monitoring under domain shifted conditions. arXiv [Preprint] (2021). http:\/\/arxiv.org\/abs\/2106.04492 (Accessed 25 November 2023)."},{"key":"e_1_3_4_65_2","doi-asserted-by":"crossref","unstructured":"A. Lopez-Lira Y. Tang Can chatGPT forecast stock price movements? Return predictability and large language models arXiv [Preprint] (2023). http:\/\/arxiv.org\/abs\/2304.07619 (Accessed 25 November 2023).","DOI":"10.2139\/ssrn.4412788"},{"key":"e_1_3_4_66_2","doi-asserted-by":"publisher","DOI":"10.1086\/689213"},{"key":"e_1_3_4_67_2","doi-asserted-by":"publisher","DOI":"10.1287\/isre.2017.0703"},{"key":"e_1_3_4_68_2","unstructured":"K. Agrawal S. Athey A. Kanodia E. Palikot Personalized recommendations in EdTech: Evidence from a randomized controlled trial. arXiv [Preprint] (2022). http:\/\/arxiv.org\/abs\/2208.13940 (Accessed 25 November 2023)."},{"key":"e_1_3_4_69_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-021-04086-x"},{"key":"e_1_3_4_70_2","doi-asserted-by":"publisher","DOI":"10.1162\/neco_a_01563"},{"key":"e_1_3_4_71_2","doi-asserted-by":"publisher","DOI":"10.1038\/s42256-022-00458-8"},{"key":"e_1_3_4_72_2","doi-asserted-by":"publisher","DOI":"10.1126\/science.aal4230"},{"key":"e_1_3_4_73_2","doi-asserted-by":"crossref","unstructured":"T. McCoy E. Pavlick T. Linzen \u201cRight for the wrong reasons: Diagnosing syntactic heuristics in natural language inference\u201d in Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics A. Korhonen D. Traum L M\u00e0rquez Eds. (Association for Computational Linguistics 2019) pp. 3428\u20133448.","DOI":"10.18653\/v1\/P19-1334"},{"key":"e_1_3_4_74_2","doi-asserted-by":"crossref","unstructured":"N. McKenna . Sources of hallucination by large language models on inference tasks. arXiv [Preprint] (2023). http:\/\/arxiv.org\/abs\/2305.14552 (Accessed 25 November 2023).","DOI":"10.18653\/v1\/2023.findings-emnlp.182"},{"key":"e_1_3_4_75_2","doi-asserted-by":"crossref","unstructured":"P. Brookins J. M. DeBacker Playing games with GPT: What can we learn about a large language model from canonical strategic games? Available at SSRN 4493398 (2023).","DOI":"10.2139\/ssrn.4493398"},{"key":"e_1_3_4_76_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 International Conference on Machine Learning A. Krause E. Brunskill K. Cho B. Engelhardt S. Sabato J. Scarlett Eds. (PMLR 2023) pp. 337\u2013371."},{"key":"e_1_3_4_77_2","first-page":"493","article-title":"Rational decision making in business organizations","volume":"69","author":"Simon H. A.","year":"1979","unstructured":"H. A. Simon, Rational decision making in business organizations. Am. Econ. Rev. 69, 493\u2013513 (1979).","journal-title":"Am. Econ. Rev."},{"key":"e_1_3_4_78_2","doi-asserted-by":"publisher","DOI":"10.1257\/aer.106.7.1577"},{"key":"e_1_3_4_79_2","unstructured":"Y. Chen T. X. Liu Y. Shan S. Zhong Data for \u201cThe emergence of economic rationality of GPT.\u201d Dropbox. https:\/\/www.dropbox.com\/scl\/fo\/572ptz57vjis5cqkczj9l\/h?rlkey=hhpsgdb6ghdzsrvj35mafdnwc&dl=0. Deposited 6 November 2023."}],"container-title":["Proceedings of the National Academy of Sciences"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/pnas.org\/doi\/pdf\/10.1073\/pnas.2316205120","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,12,12]],"date-time":"2023-12-12T13:40:10Z","timestamp":1702388410000},"score":1,"resource":{"primary":{"URL":"https:\/\/pnas.org\/doi\/10.1073\/pnas.2316205120"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,12,12]]},"references-count":79,"journal-issue":{"issue":"51","published-print":{"date-parts":[[2023,12,19]]}},"alternative-id":["10.1073\/pnas.2316205120"],"URL":"https:\/\/doi.org\/10.1073\/pnas.2316205120","relation":{},"ISSN":["0027-8424","1091-6490"],"issn-type":[{"value":"0027-8424","type":"print"},{"value":"1091-6490","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,12,12]]},"assertion":[{"value":"2023-09-22","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-11-13","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-12-12","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e2316205120"}}