Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Correspondence
  • Published:

Will generalist medical artificial intelligence be the future path for health-related natural language processing models?

Prostate Cancer and Prostatic Diseases (2023)Cite this article

Subjects

The evaluation of the effectiveness of artificial intelligence (AI) in medicine is a fascinating topic, and its application in urology has been extensively discussed in the past few months [1,2,3,4]. We read the article by Cocci et al. with great interest, which evaluated the quality and appropriateness of ChatGPT’s responses to urology-related queries [1]. As noted by the authors, natural language processing (NLP) tools such as ChatGPT currently lack sufficient appropriateness and exhibit poor quality, which align with the findings of numerous other studies [1, 2]. Undoubtedly, the rapid development of highly flexible, reusable AI models will usher in newfound capabilities in medicine. The existing models, such as ChatGPT, have showcased extraordinary capabilities due to larger datasets, increased model sizes, and advancements in model architecture. Nevertheless, these language models still fall short of fully satisfying the requirements of medical AI. Hence, the question remains: what does the future hold for the paradigm of medical AI?

Taking inspiration from the foundation models outside medicine, Moor et al. proposed a new paradigm for medical AI: generalist medical artificial intelligence (GMAI) [5]. This innovative paradigm holds tremendous potential, offering opportunities across a wide range of medical tasks. To summarize, GMAI models differentiate themselves from existing medical AI models through three crucial capabilities. Firstly, the adaptation of a GMAI model to a new task becomes effortless by simply describing the task in plain English or any other language. These models can effectively solve unfamiliar problems through dynamic task specification, eliminating the need for retraining. Secondly, GMAI models exhibit the flexibility to accept diverse input formats and generate outputs using various combinations of data modalities. In contrast to rigid multimodal models that rely on predetermined sets of input and output modalities, GMAI models empower interactive versatility. Lastly, GMAI models embody medical knowledge in a formal manner, enabling them to reason through uncharted tasks and articulate their outputs using medically accurate language. In contrast, the authors found that ChatGPT responses only demonstrated an overall accuracy of 52%, which is considered insufficient for a medical model, particularly in emergency urology cases where the accuracy was only 11.1% [1].

This is a preview of subscription content, access via your institution

Relevant articles

Open Access articles citing this article.

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

References

  1. Cocci A, Pezzoli M, Lo Re M, Russo GI, Asmundo MG, Fode M, et al. Quality of information and appropriateness of ChatGPT outputs for urology patients. Prostate Cancer Prostatic Dis. 2023;101038. https://doi.org/10.1038/s41391-023-00705-y.

  2. Davis R, Eppler M, Ayo-Ajibola O, Loh-Doyle JC, Nabhani J, Samplaski M, et al. Evaluating the effectiveness of artificial intelligence-powered large language models application in disseminating appropriate and readable health information in urology. J Urol. 2023;101097. https://doi.org/10.1097/JU.0000000000003656.

  3. Xv Y, Peng C, Wei Z, Liao F, Xiao M. Can Chat-GPT a substitute for urological resident physician in diagnosing diseases? A preliminary conclusion from an exploratory investigation. World J Urol. 2023;41:2569–71.

    Article  PubMed  Google Scholar 

  4. Deebel NA, Terlecki R. ChatGPT performance on the American Urological Association self-assessment study program and the potential influence of artificial intelligence in urologic training. Urology. 2023;177:29–33.

    Article  PubMed  Google Scholar 

  5. Moor M, Banerjee O, Abad ZSH, Krumholz HM, Leskovec J, Topol EJ, et al. Foundation models for generalist medical artificial intelligence. Nature. 2023;616:259–65.

    Article  CAS  PubMed  Google Scholar 

Download references

Acknowledgements

This article was a result of a collaborative effort involving multiple institutions. The authors would like to extend their gratitude to Prof. Y.Q. Wang from Hebei University (HBU) for providing insightful explanations about the generalist medical artificial intelligence (GMAI).

Author information

Authors and Affiliations

  1. Department of Pediatrics, Xuzhou Medical University, Xuzhou, 221004, China

    Huajie Di & Yi Wen

  2. Evidence-Based Medicine Research Center, Xuzhou Medical University, Xuzhou, 221004, China

    Huajie Di

  3. Department of Pediatric Urology, The Affiliated Xuzhou Children’s Hospital of Xuzhou Medical University, Xuzhou, 221002, China

    Huajie Di & Yi Wen

Authors
  1. Huajie Di
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Yi Wen
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

Conceptualization: HJD, YW; Writing—Original Draft: HJD; Writing—Review & Editing: HJD, YW.

Corresponding author

Correspondence to Yi Wen.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

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

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Di, H., Wen, Y. Will generalist medical artificial intelligence be the future path for health-related natural language processing models?. Prostate Cancer Prostatic Dis (2023). https://doi.org/10.1038/s41391-023-00719-6

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1038/s41391-023-00719-6

This article is cited by

Search

Advanced search

Quick links