{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,23]],"date-time":"2026-03-23T22:43:40Z","timestamp":1774305820528,"version":"3.50.1"},"reference-count":32,"publisher":"Springer Science and Business Media LLC","issue":"5","license":[{"start":{"date-parts":[[2025,8,6]],"date-time":"2025-08-06T00:00:00Z","timestamp":1754438400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,8,6]],"date-time":"2025-08-06T00:00:00Z","timestamp":1754438400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["AI Ethics"],"published-print":{"date-parts":[[2025,10]]},"abstract":"<jats:title>Abstract<\/jats:title>\n          <jats:p> Healthcare systems increasingly deploy artificial intelligence to allocate resources, including procedure authorizations that impact patient access to care. While concerns about algorithmic bias typically focus on representation of protected attributes, how AI systems approach resource-constrained decisions remains understudied. We evaluated three large language models (LLMs), ChatGPT, Claude, and DeepSeek, on their handling of simulated surgical authorization request for an identical procedure. Each model assessed 6,500 surgeon profiles while implementing a mandated 30% denial rate, mirroring real-world authorization constraints. Multivariate regression analysis quantified how each model weighted 13 standardized attributes including professional qualifications and demographic characteristics. ChatGPT assigned significantly lower authorization scores to female surgeons (-9.55 points; 95% CI: -9.98, -9.11) while Claude (+ 2.01 points; 95% CI: + 1.85, + 2.17) and DeepSeek (+ 4.03 points; 95% CI: + 3.91, + 4.15) assigned higher scores to female surgeons. Geographic biases existed, with ChatGPT heavily favoring North American surgeons (+ 18.83 points; 95% CI: + 18.00, + 19.65) and DeepSeek penalizing them (-3.95 points; 95% CI: -4.18, -3.72). In ChatGPT, demographic factors frequently outweighed clinical qualifications; geographic location impacted authorization scores more than board certification. Though all models showed high internal consistency (R<jats:sup>2<\/jats:sup> values 0.822\u20130.929), variability in prioritization of attributes resulted in divergent approval thresholds despite identical denial rates (ChatGPT: 64.6 \u00b1 21.1, Claude: 68.5 \u00b1 9.1, DeepSeek: 89.4 \u00b1 9.2). We describe a phenomenon in AI healthcare decision-making which we term \"constrained-resource divergence.\" When forced to discriminate between identical cases under resource constraints, AI systems may apply arbitrary weights that can impact patient care without clinical justification. In practice, this means patients with identical presentations may receive different authorization decisions based on which AI model their insurer deployed. Our findings raise profound questions about AI reliability for consequential healthcare decisions.<\/jats:p>","DOI":"10.1007\/s43681-025-00795-5","type":"journal-article","created":{"date-parts":[[2025,8,6]],"date-time":"2025-08-06T00:19:52Z","timestamp":1754439592000},"page":"5575-5585","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":1,"title":["The authorization lottery: contradictory AI prioritization patterns in healthcare resource allocation"],"prefix":"10.1007","volume":"5","author":[{"given":"Daniel","family":"Schneider","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0003-6933-0745","authenticated-orcid":false,"given":"Ethan","family":"Brown","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Max","family":"Ward","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Daniel","family":"Sciubba","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Sheng-Fu","family":"Lo","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"297","published-online":{"date-parts":[[2025,8,6]]},"reference":[{"key":"795_CR1","doi-asserted-by":"publisher","first-page":"149","DOI":"10.71222\/7ep7ng12","volume":"2","author":"J Zuo","year":"2025","unstructured":"Zuo, J.: AI usage case study in healthcare and health insurance sector. Econ. Manag. Innov. 2, 149\u2013158 (2025). https:\/\/doi.org\/10.71222\/7ep7ng12","journal-title":"Econ. Manag. Innov."},{"key":"795_CR2","unstructured":"Revolutionizing claims processing in the healthcare industry: the expanding role of automation and AI. In: ResearchGate (2024). https:\/\/www.researchgate.net\/publication\/388177048_Revolutionizing_Claims_Processing_in_the_Healthcare_Industry_The_Expanding_Role_of_Automation_and_AI. Accessed 13 Mar 2025"},{"key":"795_CR3","doi-asserted-by":"publisher","first-page":"100146","DOI":"10.1016\/j.cmpbup.2024.100146","volume":"5","author":"M Khalifa","year":"2024","unstructured":"Khalifa, M., Albadawy, M.: AI in diagnostic imaging: revolutionising accuracy and efficiency. Comput. Methods Programs Biomed. Update 5, 100146 (2024). https:\/\/doi.org\/10.1016\/j.cmpbup.2024.100146","journal-title":"Comput. Methods Programs Biomed. Update"},{"key":"795_CR4","doi-asserted-by":"publisher","first-page":"100148","DOI":"10.1016\/j.cmpbup.2024.100148","volume":"5","author":"M Khalifa","year":"2024","unstructured":"Khalifa, M., Albadawy, M.: Artificial intelligence for clinical prediction: exploring key domains and essential functions. Comput. Methods Programs Biomed. Update 5, 100148 (2024). https:\/\/doi.org\/10.1016\/j.cmpbup.2024.100148","journal-title":"Comput. Methods Programs Biomed. Update"},{"key":"795_CR5","doi-asserted-by":"publisher","first-page":"e2425993","DOI":"10.1001\/jamanetworkopen.2024.25993","volume":"7","author":"SW Lee","year":"2024","unstructured":"Lee, S.W., Morcos, M., Lee, D.W., Young, J.: Demographic representation of generative artificial intelligence images of physicians. JAMA Netw. Open 7, e2425993 (2024). https:\/\/doi.org\/10.1001\/jamanetworkopen.2024.25993","journal-title":"JAMA Netw. Open"},{"key":"795_CR6","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1057\/s41599-023-01999-y","volume":"10","author":"CH Chu","year":"2023","unstructured":"Chu, C.H., Donato-Woodger, S., Khan, S.S., et al.: Age-related bias and artificial intelligence: a scoping review. Humanit. Soc. Sci. Commun. 10, 1\u201317 (2023). https:\/\/doi.org\/10.1057\/s41599-023-01999-y","journal-title":"Humanit. Soc. Sci. Commun."},{"key":"795_CR7","doi-asserted-by":"publisher","first-page":"2362","DOI":"10.1016\/j.cgh.2024.09.001","volume":"22","author":"ST Amann","year":"2024","unstructured":"Amann, S.T., Shah, E.D., Wilson, L.J.: An overview of the challenging process of prior authorization: medical necessity for a better way. Clin. Gastroenterol. Hepatol. 22, 2362\u20132366 (2024). https:\/\/doi.org\/10.1016\/j.cgh.2024.09.001","journal-title":"Clin. Gastroenterol. Hepatol."},{"key":"795_CR8","doi-asserted-by":"publisher","first-page":"S65","DOI":"10.1016\/j.arth.2024.03.013","volume":"39","author":"JD Lizcano","year":"2024","unstructured":"Lizcano, J.D., Goh, G.S., Tarabichi, S., et al.: Prior authorization leads to administrative burden and delays in treatment in primary total joint arthroplasty patients. J. Arthroplasty 39, S65\u2013S70 (2024). https:\/\/doi.org\/10.1016\/j.arth.2024.03.013","journal-title":"J. Arthroplasty"},{"key":"795_CR9","doi-asserted-by":"publisher","first-page":"269","DOI":"10.1016\/j.jss.2023.03.013","volume":"288","author":"JR Merrill","year":"2023","unstructured":"Merrill, J.R., Flitcroft, M.A., Miller, T., et al.: Patterns of unnecessary insurer prior authorization denials in a complex surgical oncology practice. J. Surg. Res. 288, 269\u2013274 (2023). https:\/\/doi.org\/10.1016\/j.jss.2023.03.013","journal-title":"J. Surg. Res."},{"key":"795_CR10","unstructured":"United States Department of Health and Human Services Office of Inspector General Office of Evaluation and Inspections. High rates of prior authorization denials by some plans and limited state oversight raise concerns about access to care in Medicaid managed care. U.S. Department of Health and Human Services, Office of Inspector General, Washington, D.C. (2023)"},{"key":"795_CR11","doi-asserted-by":"publisher","first-page":"127e","DOI":"10.1097\/PRS.0000000000006968","volume":"146","author":"S Boukovalas","year":"2020","unstructured":"Boukovalas, S., Boson, A.L., Padilla, P.L., et al.: Insurance denials in reduction mammaplasty: how can we serve our patients better? Plast. Reconstr. Surg. 146, 127e (2020). https:\/\/doi.org\/10.1097\/PRS.0000000000006968","journal-title":"Plast. Reconstr. Surg."},{"key":"795_CR12","doi-asserted-by":"crossref","unstructured":"Dai, T., Abr\u00e0moff, M.D.: Incorporating artificial intelligence into healthcare workflows: models and insights. In: Tutorials in Operations Research: Advancing the Frontiers of OR\/MS: From Methodologies to Applications. INFORMS, pp. 133\u2013155 (2023)","DOI":"10.1287\/educ.2023.0257"},{"key":"795_CR13","doi-asserted-by":"publisher","first-page":"1435","DOI":"10.1007\/s10278-021-00525-3","volume":"34","author":"DJ Blezek","year":"2021","unstructured":"Blezek, D.J., Olson-Williams, L., Missert, A., Korfiatis, P.: AI integration in the clinical workflow. J. Digit. Imaging 34, 1435\u20131446 (2021). https:\/\/doi.org\/10.1007\/s10278-021-00525-3","journal-title":"J. Digit. Imaging"},{"key":"795_CR14","doi-asserted-by":"publisher","unstructured":"Shachar, C., Killelea, A., Gerke, S.: AI and health insurance prior authorization: regulators need to step up oversight. https:\/\/doi.org\/10.1377\/forefront.20240703.824037","DOI":"10.1377\/forefront.20240703.824037"},{"key":"795_CR15","doi-asserted-by":"publisher","first-page":"432","DOI":"10.3390\/a17100432","volume":"17","author":"A Lastrucci","year":"2024","unstructured":"Lastrucci, A., Pirrera, A., Lepri, G., Giansanti, D.: Algorethics in healthcare: balancing innovation and integrity in AI development. Algorithms 17, 432 (2024). https:\/\/doi.org\/10.3390\/a17100432","journal-title":"Algorithms"},{"key":"795_CR16","doi-asserted-by":"publisher","first-page":"e0000022","DOI":"10.1371\/journal.pdig.0000022","volume":"1","author":"LA Celi","year":"2022","unstructured":"Celi, L.A., Cellini, J., Charpignon, M.-L., et al.: Sources of bias in artificial intelligence that perpetuate healthcare disparities\u2014A global review. PLOS Digit. Health 1, e0000022 (2022). https:\/\/doi.org\/10.1371\/journal.pdig.0000022","journal-title":"PLOS Digit. Health"},{"key":"795_CR17","doi-asserted-by":"publisher","first-page":"407","DOI":"10.1007\/s11673-022-10200-z","volume":"19","author":"J Delgado","year":"2022","unstructured":"Delgado, J., de Manuel, A., Parra, I., et al.: Bias in algorithms of AI systems developed for COVID-19: a scoping review. J. Bioethical Inq. 19, 407\u2013419 (2022). https:\/\/doi.org\/10.1007\/s11673-022-10200-z","journal-title":"J. Bioethical Inq."},{"key":"795_CR18","unstructured":"Specification gaming: the flip side of AI ingenuity. In: Google Deep (2025). https:\/\/deepmind.google\/discover\/blog\/specification-gaming-the-flip-side-of-ai-ingenuity\/. Accessed 13 Mar 2025"},{"key":"795_CR19","unstructured":"Pan, A., Bhatia, K., Steinhardt, J.: The effects of reward misspecification: mapping and mitigating misaligned models (2022)"},{"key":"795_CR20","doi-asserted-by":"publisher","first-page":"102779","DOI":"10.1016\/j.artmed.2024.102779","volume":"149","author":"S Aminizadeh","year":"2024","unstructured":"Aminizadeh, S., Heidari, A., Dehghan, M., et al.: Opportunities and challenges of artificial intelligence and distributed systems to improve the quality of healthcare service. Artif. Intell. Med. 149, 102779 (2024). https:\/\/doi.org\/10.1016\/j.artmed.2024.102779","journal-title":"Artif. Intell. Med."},{"key":"795_CR21","unstructured":"AI Lawsuits Against Insurers Signal Wave of Health Litigation (2024). https:\/\/news.bloomberglaw.com\/health-law-and-business\/ai-lawsuits-against-insurers-signal-wave-of-health-litigation. Accessed 13 Mar 2025"},{"key":"795_CR22","unstructured":"Byrd, J., Lipton, Z.: What is the effect of importance weighting in deep learning? In: Proceedings of the 36th International Conference on Machine Learning. PMLR, pp 872\u2013881 (2019)"},{"key":"795_CR23","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1007\/978-3-540-27774-3_7","volume-title":"Modeling Decisions for Artificial Intelligence","author":"T Nakashima","year":"2004","unstructured":"Nakashima, T., Ishibuchi, H., Bargiela, A.: A study on weighting training patterns for fuzzy rule-based classification systems. In: Torra, V., Narukawa, Y. (eds.) Modeling Decisions for Artificial Intelligence, pp. 60\u201369. Springer, Berlin (2004)"},{"key":"795_CR24","doi-asserted-by":"crossref","unstructured":"Sp\u00e4th, H., Daw, Z.: Towards detecting unintended behaviors in machine learning algorithms. In: 2024 AIAA DATC\/IEEE 43rd Digital Avionics Systems Conference (DASC), pp. 1\u201310 (2024)","DOI":"10.1109\/DASC62030.2024.10749477"},{"key":"795_CR25","first-page":"3650","volume-title":"Advances in Neural Information Processing Systems","author":"P Hase","year":"2021","unstructured":"Hase, P., Xie, H., Bansal, M.: The out-of-distribution problem in explainability and search methods for feature importance explanations. In: Advances in Neural Information Processing Systems, pp. 3650\u20133666. Curran Associates Inc., Red Hook (2021)"},{"key":"795_CR26","doi-asserted-by":"publisher","first-page":"562","DOI":"10.3390\/healthcare12050562","volume":"12","author":"K Palaniappan","year":"2024","unstructured":"Palaniappan, K., Lin, E.Y.T., Vogel, S.: Global regulatory frameworks for the use of artificial intelligence (AI) in the healthcare services sector. Healthcare 12, 562 (2024). https:\/\/doi.org\/10.3390\/healthcare12050562","journal-title":"Healthcare"},{"key":"795_CR27","unstructured":"Health C for D and R. Software as a medical device (SaMD). In: FDA (2024). https:\/\/www.fda.gov\/medical-devices\/digital-health-center-excellence\/software-medical-device-samd. Accessed 13 Mar 2025"},{"key":"795_CR28","unstructured":"AI Act| Shaping Europe\u2019s digital future (2025). https:\/\/digital-strategy.ec.europa.eu\/en\/policies\/regulatory-framework-ai. Accessed 13 Mar 2025"},{"key":"795_CR29","doi-asserted-by":"publisher","DOI":"10.5281\/zenodo.14598858","author":"P Tupsakhare","year":"2025","unstructured":"Tupsakhare, P., Kulkarni, T.: Healthcare technology testing: an AI-driven approach. IJAIDR J. Adv. Dev. Res. (2025). https:\/\/doi.org\/10.5281\/zenodo.14598858"},{"key":"795_CR30","doi-asserted-by":"publisher","first-page":"117519","DOI":"10.1016\/j.cca.2023.117519","volume":"548","author":"S Polevikov","year":"2023","unstructured":"Polevikov, S.: Advancing AI in healthcare: a comprehensive review of best practices. Clin. Chim. Acta 548, 117519 (2023). https:\/\/doi.org\/10.1016\/j.cca.2023.117519","journal-title":"Clin. Chim. Acta"},{"key":"795_CR31","doi-asserted-by":"publisher","first-page":"e100444","DOI":"10.1136\/bmjhci-2021-100444","volume":"28","author":"S Reddy","year":"2021","unstructured":"Reddy, S., Rogers, W., Makinen, V.-P., et al.: Evaluation framework to guide implementation of AI systems into healthcare settings. BMJ Health Care Inform. 28, e100444 (2021). https:\/\/doi.org\/10.1136\/bmjhci-2021-100444","journal-title":"BMJ Health Care Inform."},{"key":"795_CR32","doi-asserted-by":"publisher","first-page":"e13087","DOI":"10.1111\/ecc.13087","volume":"28","author":"L Johnston","year":"2019","unstructured":"Johnston, L., Young, J., Campbell, K.: The implementation and impact of Holistic Needs Assessments for people affected by cancer: a systematic review and thematic synthesis of the literature. Eur. J. Cancer Care 28, e13087 (2019). https:\/\/doi.org\/10.1111\/ecc.13087","journal-title":"Eur. J. Cancer Care"}],"container-title":["AI and Ethics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s43681-025-00795-5.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s43681-025-00795-5\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s43681-025-00795-5.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,19]],"date-time":"2025-09-19T08:34:48Z","timestamp":1758270888000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s43681-025-00795-5"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,8,6]]},"references-count":32,"journal-issue":{"issue":"5","published-print":{"date-parts":[[2025,10]]}},"alternative-id":["795"],"URL":"https:\/\/doi.org\/10.1007\/s43681-025-00795-5","relation":{"references":[{"id-type":"doi","id":"10.5281\/zenodo.14598858","asserted-by":"subject"}]},"ISSN":["2730-5953","2730-5961"],"issn-type":[{"value":"2730-5953","type":"print"},{"value":"2730-5961","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,8,6]]},"assertion":[{"value":"28 April 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"6 July 2025","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"6 August 2025","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no competing interests.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Conflict of interest"}},{"value":"This study was considered exempt from IRB approval due to its use of synthetic, anonymized data.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethical approval"}}]}}