Whose doctor does the AI recommend? An algorithm audit of LLMs in physician

A prespecified randomized algorithm audit of seven large language models (six open-weight and gpt-4o-mini) found that reputation signals dominate physician recommendations, with raising a rating from 3.9 to 4.7 increasing choice probability by 31.4 percentage points and raising the fee from $90 to $190 lowering it by 20.0 percentage points. The audit, which analyzed 40,068 scored responses across 3,024 choice sets, also found demographic parity is rejected in the opposite direction of human audit studies: female-signaled names gain 2.5 pp and Hispanic-, South-Asian-, and Black-signaled names gain 1.3-2.9 pp over White-signaled names, yet models mentioned gender or ethnicity in at most 0.03% of stated reasons, making these effects invisible to self-report. One reasoning model failed the prespecified auditability gate, and the authors argue recurring behavioral audit is the monitoring technology fit for purpose.

Computer Science Computers and Society Submitted on 14 Aug 2026 Title:Whose doctor does the AI recommend? An algorithm audit of reputation and demographic signals in large language model-assisted physician choice View PDF /pdf/2608.14399 HTML experimental https://arxiv.org/html/2608.14399v1 Abstract:Patients increasingly ask large language model LLM assistants which doctor to see, making these systems AI infomediaries: algorithms that intermediate one person's choice among other people and thereby decide, silently and at scale, which physicians become visible. We report a prespecified randomized algorithm audit of what causally moves those recommendations. Seven models six open-weight; gpt-4o-mini each chose among five synthetic family-medicine physician cards whose attributes were independently randomized across 3,024 choice sets, three patient personas, nine prompt paraphrases and nine experimental arms, yielding 40,068 scored responses; gender and ethnicity were signaled through names following correspondence-audit methodology. Reputation signals dominate: raising a rating from 3.9 to 4.7 increases choice probability by 31.4 percentage points pp , and raising the fee from $90 to $190 lowers it by 20.0 pp. Demographic parity is rejected, but not in the direction human audit studies predict: female-signaled names gain 2.5 pp, and Hispanic-, South-Asian- and Black-signaled names gain 1.3-2.9 pp over White-signaled names, tilts worth $7-$14 per visit in fee-equivalent terms, and a content-free first-listed position is worth $11. Yet models mentioned gender or ethnicity in at most 0.03% of their stated reasons and abstained in 0.39% of trials, so these effects are invisible in the models' own explanations, and transparency obligations relying on model self-report would not detect them. One reasoning model failed the prespecified auditability gate outright. The frozen design makes the audit repeatable: any new model can be assessed against identical stimuli, making recurring behavioural audit, rather than self-reported explanation, the monitoring technology fit for purpose. Submission history From: Mirza Samad Ahmed Baig view email /show-email/57a95a53/2608.14399 v1 Fri, 14 Aug 2026 15:39:10 UTC 86 KB Current browse context: cs.CY References & Citations Loading... Bibliographic and Citation Tools Bibliographic Explorer What is the Explorer? https://info.arxiv.org/labs/showcase.html arxiv-bibliographic-explorer Connected Papers What is Connected Papers? https://www.connectedpapers.com/about Litmaps What is Litmaps? https://www.litmaps.co/ scite Smart Citations What are Smart Citations? https://www.scite.ai/ Code, Data and Media Associated with this Article alphaXiv What is alphaXiv? https://alphaxiv.org/ CatalyzeX Code Finder for Papers What is CatalyzeX? https://www.catalyzex.com DagsHub What is DagsHub? https://dagshub.com/ Gotit.pub What is GotitPub? http://gotit.pub/faq Hugging Face What is Huggingface? https://huggingface.co/huggingface ScienceCast What is ScienceCast? https://sciencecast.org/welcome Demos Recommenders and Search Tools Influence Flower What are Influence Flowers? https://influencemap.cmlab.dev/ CORE Recommender What is CORE? https://core.ac.uk/services/recommender arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs https://info.arxiv.org/labs/index.html .