跳到正文
原文
Nature 健康护理主题·· 2026-10-05AI 评分44

数字健康中大型语言模型的负责任使用:基于公平优先的治理框架

Responsible use of large language models in digital health: an equity-first governance framework

AI 导读

针对三款开源大模型(LLaMA-3.2、Mistral-7B、DeepSeek-r1:8B)的审计显示,明确医疗免责声明在 1350 条响应中占比均低于 0.2%。危机指导语在 LLaMA 和 Mistral 中约 99.6% 出现,但在 DeepSeek 中仅 71.3%,且截断至 180 词时该比例大幅降至 30.0%。结果提示技术流畅性不保证公平充分性,需结合具体治理行动。

正文

Scientific Reports (2026) Cite this article

We’re sharing this article early to provide faster access to peer-reviewed, accepted research. It is citable and carries a permanent DOI. This version is subject to further edits and will be replaced automatically by the final Version of Record. All legal disclaimers apply.

Abstract

Large language models (LLMs) are being rapidly deployed for digital health, yet their equity impacts remain poorly characterized. We present an equity-first audit framework anchored in the first nations mental wellness continuum framework (FNMWCF) and apply it to three open-source LLMs (LLaMA-3.2:latest, Mistral-7B, and DeepSeek-r1:8B). Using three FNMWCF-informed persona prompts derived from the Canadian Community Health Survey (CCHS), we generated 150 stochastic responses per persona from each model (temperature \(=0.7\)), yielding 450 responses per model and 1350 responses in total. We analysed the full corpus with automated, reproducible text metrics and separately conducted a small, exploratory single-rater human pilot on a 0–3 rubric. Across all 1,350 responses, lexical cultural markers were near-universal (100%), but explicit safety scaffolding varied: assessed on full responses, crisis-guidance language appeared in \(\approx\)99.6% of LLaMA-3.2 and Mistral-7B responses and in 71.3% of DeepSeek-r1:8B responses, whereas an explicit medical-disclaimer/clinician-referral marker (a “not medical advice” statement or a recommendation to consult a doctor/clinician/provider) was almost entirely absent (\(\le\)0.2%). A sensitivity analysis showed that the 180-word display trim materially reduced the measured DeepSeek-r1:8B crisis prevalence (to 30.0%; a 41-point drop) while barely affecting the other models, indicating that trimming can remove late-appearing safety content. Because the personas describe moderate-severity, non-emergency wellness scenarios and the prompt explicitly requests crisis and urgent-care guidance, the high crisis-guidance prevalence may reflect both prompt adherence and a more conservative safety tendency in some models rather than uniformly safer behaviour; marker prevalence alone cannot establish whether crisis messaging is calibrated to scenario severity. Readability estimates (Flesch–Kincaid grade \(\approx\)10–11) exceeded the grade 6–8 range recommended for patient-facing materials. FNMWCF theme terms appeared frequently, but such lexical presence primarily reflects adherence to the FNMWCF-informed prompt and is a proxy only; it does not establish cultural adequacy. The human rubric pilot (13 rated DeepSeek responses, one rater) is reported as exploratory; it does not support a model comparison, and inter-rater reliability could not be computed with the available data. Within the evaluated models, prompts, and experimental conditions, our findings suggest that technical fluency is no guarantee of equitable adequacy. We provide a reproducible equity-first audit pipeline and an operational governance framework linking each empirical signal to a concrete oversight action.

Similar content being viewed by others

Acknowledgements

The authors thank the Artificial Intelligence and Mathematical Modeling Lab (AIMMlab) at the University of Toronto and members of the Africa-Canada Artificial Intelligence and Data Innovation Consortium (ACADIC) for their input and collaboration.

Funding

This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant (RGPIN-2022-04559). J.D.K. acknowledges additional support from the NSERC Discovery Launch Supplement (DGECR-2022-00454), the SSHRC New Frontiers in Research Fund – Exploratory (NFRFE-2021-00879), Canada’s International Development Research Centre (IDRC) (109981), IDRC and the Foreign, Commonwealth and Development Office (FCDO) (110554-001 and 110693-001), and the Canada Research Chair in Community-Oriented Artificial Intelligence and Mathematical Modelling of Infectious Diseases (CRC-2023-00234).

Author information

Authors and Affiliations

  1. Decentralized Cybersecurity and Artificial Intelligence Lab (DCAILab), Department of Computer Science, University of Regina, Regina, SK, Canada

    Abbas Yazdinejad

  2. Artificial Intelligence and Mathematical Modeling Lab (AIMMlab), Dalla Lana School of Public Health, University of Toronto, Toronto, ON, Canada

    Jude Kong

  3. Institute of Health Policy, Management and Evaluation (IHPME), University of Toronto, Toronto, ON, Canada

    Jude Kong

  4. Department of Mathematics, University of Toronto, Bahen Centre for Information Technology, Toronto, ON, Canada

    Jude Kong

  5. Africa-Canada Artificial Intelligence and Data Innovation Consortium (ACADIC), Toronto, ON, Canada

    Abbas Yazdinejad & Jude Kong

  6. Global South Artificial Intelligence for Pandemic and Epidemic Preparedness and Response Network (AI4PEP), Toronto, ON, Canada

    Abbas Yazdinejad & Jude Kong

Authors

  1. Abbas Yazdinejad
  2. Jude Kong

Corresponding authors

Correspondence to Abbas Yazdinejad or Jude Kong.

Ethics declarations

Competing interests

The authors declare no competing interests.

Use of AI-assisted tools

During the preparation of this manuscript, the authors used AI-assisted tools to support language and copy editing, code review and reproducibility checking, regeneration of figures and tables from the underlying data, consistency checks across the manuscript, and drafting of the point-by-point responses to reviewers. All analyses, results, interpretations, and conclusions were reviewed and verified by the authors, who take full responsibility for the accuracy, originality, and integrity of the work.

Additional information

Publisher's note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Yazdinejad, A., Kong, J. Responsible use of large language models in digital health: an equity-first governance framework. Sci Rep (2026). https://doi.org/10.1038/s41598-026-72535-6

Download citation

  • Received: 19 June 2026

  • Accepted: 16 September 2026

  • Published: 05 October 2026

  • DOI: https://doi.org/10.1038/s41598-026-72535-6

来源:Nature 健康护理主题 · nature.com