{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T16:44:57Z","timestamp":1791823497687,"version":"4.3.4"},"posted":{"date-parts":[[2026]]},"group-title":"SSRN","reference-count":73,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/www.uspto.gov\/ip-policy\/copyright-policy\/copyright-basics"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"<jats:p>Large language models can improve investors' understanding of corporate disclosures, but reliance on common models can also make their errors more correlated. I study this trade-off using the public launch of ChatGPT and two pre-launch properties of corporate filings, readability and uncertainty. After the launch, order flow in hard-to-read firms becomes more one-sided relative to comparable firms, in both retail and non-retail trading. Relative to firms in the easiest-to-read decile, one-sidedness in the hardest-to-read decile rises by $6.4$ percent of the sample mean. Price discovery also improves: for a firm one standard deviation harder to read, the share of the price impact of order flow that survives the following quarter rises from about $60$ to about $80$ percent, and earnings news enters prices faster. Non-retail order flow in hard-to-read stocks also moves more with the rest of the market. Greater disclosure uncertainty predicts less synchronized trading relative to other firms and no improvement in price discovery. To interpret these patterns, I develop a Kyle-style model in which investors can delegate reading to a common model that decodes better but introduces a shared error. Because trading on information that others also hold is less profitable, delegation is a strategic substitute, and crowding limits adoption. Greater synchronization can therefore coexist with better price discovery.<\/jats:p>","DOI":"10.2139\/ssrn.7538559","type":"posted-content","created":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T02:56:57Z","timestamp":1790650617000},"source":"Crossref","is-referenced-by-count":0,"title":["Large Language Models, Price Discovery, and Synchronized Trading"],"prefix":"10.2139","author":[{"given":"Jiawei","family":"Wang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","reference":[{"issue":"2","key":"ref1","doi-asserted-by":"crossref","first-page":"400","DOI":"10.1016\/0022-0531(86)90052-9","article-title":"A monopolistic market for information","volume":"39","author":"A R Admati","year":"1986","journal-title":"Journal of Economic Theory"},{"issue":"1","key":"ref2","doi-asserted-by":"crossref","first-page":"3","DOI":"10.1093\/rfs\/1.1.3","article-title":"A theory of intraday patterns: Volume and price variability","volume":"1","author":"A R Admati","year":"1988","journal-title":"The Review of Financial Studies"},{"key":"ref3","article-title":"Artificial intelligence in the financial system","author":"Bank","year":"2025","journal-title":"Bank of England, Financial Policy Committee"},{"key":"ref4","article-title":"Unusual financial communication: ChatGPT, earnings calls, and financial markets","author":"L Beckmann","year":"2024","journal-title":"Olin Business School Center for Finance & Accounting Research"},{"key":"ref5","author":"J Bertomeu","year":"2023","journal-title":"Capital market consequences of generative AI: Early evidence from the ban of ChatGPT in Italy. Working paper"},{"issue":"1","key":"ref6","doi-asserted-by":"crossref","DOI":"10.1016\/j.jacceco.2025.101782","article-title":"The impact of generative AI on information processing: Evidence from the ban of ChatGPT in Italy","volume":"80","author":"J Bertomeu","year":"2025","journal-title":"Journal of Accounting and Economics"},{"key":"ref7","article-title":"Generative AI and investor processing of financial information","author":"E Blankespoor","year":"2026","journal-title":"Journal of Accounting and Economics. Forthcoming"},{"issue":"3","key":"ref8","doi-asserted-by":"crossref","first-page":"1189","DOI":"10.1111\/1475-679x.70050","article-title":"Generative AI in financial reporting","volume":"64","author":"E Blankespoor","year":"2026","journal-title":"Journal of Accounting Research"},{"key":"ref9","first-page":"233","volume":"16","author":"R J Bloomfield","year":"2002","journal-title":"The \"incomplete revelation hypothesis\" and financial reporting"},{"issue":"5","key":"ref10","doi-asserted-by":"crossref","first-page":"2249","DOI":"10.1111\/jofi.13033","article-title":"Tracking retail investor activity","volume":"76","author":"E Boehmer","year":"2021","journal-title":"The Journal of Finance"},{"key":"ref11","author":"J L Campbell","year":"2026","journal-title":"Generative artificial intelligence and retail investors' processing of earnings news"},{"issue":"9","key":"ref12","doi-asserted-by":"crossref","first-page":"3603","DOI":"10.1093\/rfs\/hhad021","article-title":"How to talk when a machine is listening: Corporate disclosure in the age of AI","volume":"36","author":"S Cao","year":"2023","journal-title":"The Review of Financial Studies"},{"key":"ref13","author":"\u00c1 Cartea","year":"2026","journal-title":"AI bubbles with large language models"},{"issue":"3","key":"ref14","doi-asserted-by":"crossref","first-page":"1287","DOI":"10.1111\/1475-679x.70063","article-title":"AI democratization and trading inequality","volume":"64","author":"A Chang","year":"2026","journal-title":"Journal of Accounting Research"},{"key":"ref15","author":"J Chen","year":"2025","journal-title":"ChatGPT and DeepSeek: Can they predict the stock market and macroeconomy? arXiv preprint"},{"key":"ref16","author":"S Chen","year":"2025","journal-title":"Wisdom or whims? decoding retail strategies with social media and AI"},{"key":"ref17","doi-asserted-by":"crossref","DOI":"10.1016\/j.jacceco.2025.101821","article-title":"Does generative AI facilitate investor trading? early evidence from ChatGPT outages","volume":"80","author":"Q Cheng","year":"2025","journal-title":"Journal of Accounting and Economics"},{"issue":"1","key":"ref18","doi-asserted-by":"crossref","first-page":"3","DOI":"10.1016\/S0304-405X(99)00057-4","article-title":"Commonality in liquidity","volume":"56","author":"T Chordia","year":"2000","journal-title":"Journal of Financial Economics"},{"issue":"3","key":"ref19","doi-asserted-by":"crossref","first-page":"485","DOI":"10.1016\/S0304-405X(03)00175-2","article-title":"Order imbalance and individual stock returns: Theory and evidence","volume":"72","author":"T Chordia","year":"2004","journal-title":"Journal of Financial Economics"},{"issue":"3","key":"ref20","doi-asserted-by":"crossref","first-page":"1371","DOI":"10.1111\/jofi.12885","article-title":"Lazy prices","volume":"75","author":"L Cohen","year":"2020","journal-title":"The Journal of Finance"},{"key":"ref21","doi-asserted-by":"crossref","DOI":"10.3386\/w34054","author":"W W Dou","year":"2025","journal-title":"AI-powered trading, algorithmic collusion, and price efficiency"},{"key":"ref22","author":"A I Epoch","year":"2026","journal-title":"Data on AI models. Notable AI Models database, snapshot downloaded"},{"key":"ref23","author":"O Even-Tov","year":"2025","journal-title":"The effect of AI on retail investor behavior: Evidence from account-level trading data"},{"issue":"1","key":"ref24","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/j.jfineco.2014.10.010","article-title":"A five-factor asset pricing model","volume":"116","author":"E F Fama","year":"2015","journal-title":"Journal of financial economics"},{"key":"ref25","year":"2024","journal-title":"The financial stability implications of artificial intelligence"},{"issue":"4","key":"ref26","doi-asserted-by":"crossref","first-page":"1367","DOI":"10.1093\/rfs\/hhz100","article-title":"Informing the market: The effect of modern information technologies on information production","volume":"33","author":"M Gao","year":"2020","journal-title":"The Review of Financial Studies"},{"issue":"3","key":"ref27","doi-asserted-by":"crossref","first-page":"535","DOI":"10.1257\/jel.20181020","article-title":"Text as data","volume":"57","author":"M Gentzkow","year":"2019","journal-title":"Journal of Economic Literature"},{"issue":"3","key":"ref28","doi-asserted-by":"crossref","first-page":"617","DOI":"10.1111\/j.1540-6261.1988.tb04594.x","article-title":"Liquidity and market structure","volume":"43","author":"S J Grossman","year":"1988","journal-title":"The Journal of Finance"},{"issue":"3","key":"ref29","first-page":"393","article-title":"On the impossibility of informationally efficient markets","volume":"70","author":"S J Grossman","year":"1980","journal-title":"The American economic review"},{"key":"ref30","author":"A L Hansen","year":"2024","journal-title":"Can ChatGPT decipher Fedspeak? Available at SSRN 4399406"},{"issue":"1","key":"ref31","doi-asserted-by":"crossref","first-page":"179","DOI":"10.1111\/j.1540-6261.1991.tb03749.x","article-title":"Measuring the information content of stock trades","volume":"46","author":"J Hasbrouck","year":"1991","journal-title":"The Journal of Finance"},{"issue":"3","key":"ref32","doi-asserted-by":"crossref","first-page":"383","DOI":"10.1016\/S0304-405X(00)00091-X","article-title":"Common factors in prices, order flows, and liquidity","volume":"59","author":"J Hasbrouck","year":"2001","journal-title":"Journal of Financial Economics"},{"issue":"3","key":"ref33","doi-asserted-by":"crossref","first-page":"405","DOI":"10.1016\/j.jfineco.2014.08.001","article-title":"Price pressures","volume":"114","author":"T Hendershott","year":"2014","journal-title":"Journal of Financial Economics"},{"key":"ref34","doi-asserted-by":"crossref","DOI":"10.3386\/w34807","volume":"34807","author":"D Hirshleifer","year":"2026","journal-title":"AI, opinion ecosystems, and finance"},{"issue":"1","key":"ref35","first-page":"247","article-title":"Long-lived private information and imperfect competition","volume":"47","author":"C W Holden","year":"1992","journal-title":"The Journal of Finance"},{"key":"ref36","article-title":"Financial services in the era of generative AI: Facilitating responsible adoption","volume":"1","year":"2025","journal-title":"HKIMR Applied Research Report"},{"issue":"2","key":"ref37","doi-asserted-by":"crossref","first-page":"806","DOI":"10.1111\/1911-3846.12832","article-title":"FinBERT: A large language model for extracting information from financial text","volume":"40","author":"A H Huang","year":"2023","journal-title":"Contemporary Accounting Research"},{"key":"ref38","author":"L Huang","year":"2024","journal-title":"Viral communication in financial markets: Meme images and market efficiency"},{"key":"ref39","author":"W Huang","year":"2026","journal-title":"AI \"errors"},{"key":"ref40","year":"2024","journal-title":"Global financial stability report"},{"issue":"4","key":"ref41","first-page":"1193","article-title":"Does common analyst coverage explain excess comovement? Journal of Financial and Quantitative","volume":"51","author":"R D Israelsen","year":"2016","journal-title":"Analysis"},{"key":"ref42","doi-asserted-by":"crossref","DOI":"10.3386\/w32161","author":"M Jha","year":"2024","journal-title":"ChatGPT and corporate policies"},{"key":"ref43","doi-asserted-by":"crossref","first-page":"160","DOI":"10.1145\/3618260.3649777","article-title":"Calibrated language models must hallucinate","author":"A T Kalai","year":"2024","journal-title":"Proceedings of the 56th Annual ACM Symposium on Theory of Computing (STOC)"},{"issue":"4","key":"ref44","doi-asserted-by":"crossref","first-page":"831","DOI":"10.1086\/262005","article-title":"Differential interpretation of public signals and trade in speculative markets","volume":"103","author":"E Kandel","year":"1995","journal-title":"Journal of Political Economy"},{"key":"ref45","first-page":"15696","article-title":"Large language models struggle to learn long-tail knowledge","author":"N Kandpal","year":"2023","journal-title":"Proceedings of the 40th International Conference on Machine Learning (ICML)"},{"key":"ref46","author":"H Kang","year":"2023","journal-title":"Deficiency of large language models in finance: An empirical examination of hallucination"},{"key":"ref47","author":"A Kim","year":"2023","journal-title":"Bloated disclosures: can ChatGPT help investors process information? arXiv preprint"},{"key":"ref48","first-page":"30038","article-title":"Correlated errors in large language models","volume":"267","author":"E Kim","year":"2025","journal-title":"Proceedings of the 42nd International Conference on Machine Learning"},{"issue":"1-2","key":"ref49","doi-asserted-by":"crossref","first-page":"41","DOI":"10.1016\/0165-4101(94)90004-3","article-title":"Market liquidity and volume around earnings announcements","volume":"17","author":"O Kim","year":"1994","journal-title":"Journal of Accounting and Economics"},{"key":"ref50","doi-asserted-by":"crossref","first-page":"1315","DOI":"10.2307\/1913210","article-title":"Continuous auctions and insider trading","author":"A S Kyle","year":"1985","journal-title":"Econometrica: Journal of the Econometric Society"},{"issue":"4","key":"ref51","doi-asserted-by":"crossref","first-page":"1119","DOI":"10.3982\/ECTA12635","article-title":"Strategic trading in informationally complex environments","volume":"86","author":"N S Lambert","year":"2018","journal-title":"Econometrica"},{"issue":"2","key":"ref52","doi-asserted-by":"crossref","first-page":"733","DOI":"10.1111\/j.1540-6261.1991.tb02683.x","article-title":"Inferring trade direction from intraday data","volume":"46","author":"C M C Lee","year":"1991","journal-title":"The Journal of Finance"},{"key":"ref53","author":"G Leitner","year":"2024","journal-title":"The rise of artificial intelligence: Benefits and risks for financial stability. Financial Stability Review special feature May 2024"},{"issue":"2-3","key":"ref54","doi-asserted-by":"crossref","first-page":"221","DOI":"10.1016\/j.jacceco.2008.02.003","article-title":"Annual report readability, current earnings, and earnings persistence","volume":"45","author":"F Li","year":"2008","journal-title":"Journal of Accounting and Economics"},{"issue":"1","key":"ref55","doi-asserted-by":"crossref","first-page":"253","DOI":"10.1093\/rfs\/hhaf081","article-title":"Dissecting corporate culture using generative AI","volume":"39","author":"K Li","year":"2026","journal-title":"The Review of financial studies"},{"key":"ref56","doi-asserted-by":"crossref","first-page":"157","DOI":"10.1162\/tacl_a_00638","article-title":"Lost in the middle: How language models use long contexts","volume":"12","author":"N F Liu","year":"2024","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"ref57","article-title":"Can ChatGPT forecast stock price movements? return predictability and large language models","author":"A Lopez-Lira","year":"2026","journal-title":"Journal of Financial Economics. Forthcoming"},{"issue":"1","key":"ref58","doi-asserted-by":"crossref","first-page":"35","DOI":"10.1111\/j.1540-6261.2010.01625.x","article-title":"When is a liability not a liability? textual analysis, dictionaries, and 10-Ks","volume":"66","author":"T Loughran","year":"2011","journal-title":"The Journal of Finance"},{"issue":"4","key":"ref59","doi-asserted-by":"crossref","first-page":"1643","DOI":"10.1111\/jofi.12162","article-title":"Measuring readability in financial disclosures","volume":"69","author":"T Loughran","year":"2014","journal-title":"The Journal of Finance"},{"issue":"4","key":"ref60","doi-asserted-by":"crossref","first-page":"1187","DOI":"10.1111\/1475-679X.12123","article-title":"Textual analysis in accounting and finance: A survey","volume":"54","author":"T Loughran","year":"2016","journal-title":"Journal of Accounting Research"},{"key":"ref61","author":"F Lu","year":"2023","journal-title":"ChatGPT, generative AI, and investment advisory"},{"key":"ref62","author":"R Lu","year":"2026","journal-title":"Do LLMs make markets more efficient? Working paper"},{"issue":"6","key":"ref63","doi-asserted-by":"crossref","first-page":"2107","DOI":"10.2308\/accr.00000001","article-title":"The effects of reporting complexity on small and large investor trading","volume":"85","author":"B P Miller","year":"2010","journal-title":"The Accounting Review"},{"issue":"1-2","key":"ref64","doi-asserted-by":"crossref","first-page":"215","DOI":"10.1016\/S0304-405X(00)00071-4","article-title":"The information content of stock markets: Why do emerging markets have synchronous stock price movements","volume":"58","author":"R Morck","year":"2000","journal-title":"Journal of Financial Economics"},{"issue":"3","key":"ref65","first-page":"465","article-title":"Herd behavior and investment","volume":"80","author":"D S Scharfstein","year":"1990","journal-title":"American Economic Review"},{"key":"ref66","author":"M Shaffer","year":"2024","journal-title":"Scaling core earnings measurement with large language models"},{"key":"ref67","author":"J Sheng","year":"2026","journal-title":"Generative AI and asset management. Review of Financial Studies. Forthcoming"},{"key":"ref68","author":"J Stanco","year":"2025","journal-title":"Generative AI and the transmission of public information in the stock market: Evidence from the release of ChatGPT"},{"issue":"4","key":"ref69","doi-asserted-by":"crossref","first-page":"1517","DOI":"10.1111\/j.1540-6261.2009.01472.x","article-title":"Presidential address: Sophisticated investors and market efficiency","volume":"64","author":"J C Stein","year":"2009","journal-title":"The Journal of Finance"},{"issue":"3","key":"ref70","doi-asserted-by":"crossref","first-page":"823","DOI":"10.1111\/j.1467-937X.2006.00397.x","article-title":"Information markets and the comovement of asset prices","volume":"73","author":"L L Veldkamp","year":"2006","journal-title":"Review of Economic Studies"},{"key":"ref71","article-title":"How AI agents mislead investors: Early evidence from automated stock recommendations","author":"Y Xu","year":"2025","journal-title":"SSRN 5651130 . Table A3: Frontier model releases used in N (m)"},{"key":"ref72","article-title":"Ultra Google DeepMind 2024-02-15 Gemini 1.5 Pro Google DeepMind","journal-title":"Date Model"},{"key":"ref73","journal-title":"Anthropic Notes: A release enters the calendar if it meets four criteria. First, the model is publicly available at launch through an interface that investors can use to analyze a filing. This excludes research announcements, benchmark-only systems, and models without public access. Second, the developer is at the capability frontier at the time of release. Third, the release introduces the developer's leading general-purpose model, excluding size variants, distilled or lower-cost versions, and maintenance updates"}],"container-title":[],"original-title":[],"deposited":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T15:50:14Z","timestamp":1791820214000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ssrn.com\/abstract=7538559"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":73,"URL":"https:\/\/doi.org\/10.2139\/ssrn.7538559","relation":{},"subject":[],"published":{"date-parts":[[2026]]},"subtype":"preprint"}}