IDEAS home Printed from https://ideas.repec.org/a/nat/natcom/v12y2021i1d10.1038_s41467-021-27577-x.html
   My bibliography  Save this article

A machine and human reader study on AI diagnosis model safety under attacks of adversarial images

Author

Listed:
  • Qianwei Zhou

    (University of Pittsburgh
    Zhejiang University of Technology
    Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province)

  • Margarita Zuley

    (University of Pittsburgh
    University of Pittsburgh Medical Center)

  • Yuan Guo

    (University of Pittsburgh
    Guangzhou First People’s Hospital, School of Medicine, South China University of Technology)

  • Lu Yang

    (University of Pittsburgh
    Chongqing Key Laboratory of Translational Research for Cancer Metastasis and Individualized Treatment, Chongqing University Cancer Hospital)

  • Bronwyn Nair

    (University of Pittsburgh
    University of Pittsburgh Medical Center)

  • Adrienne Vargo

    (University of Pittsburgh
    University of Pittsburgh Medical Center)

  • Suzanne Ghannam

    (University of Pittsburgh
    University of Pittsburgh Medical Center)

  • Dooman Arefan

    (University of Pittsburgh)

  • Shandong Wu

    (University of Pittsburgh
    University of Pittsburgh
    University of Pittsburgh
    Intelligent Systems Program, University of Pittsburgh)

Abstract

While active efforts are advancing medical artificial intelligence (AI) model development and clinical translation, safety issues of the AI models emerge, but little research has been done. We perform a study to investigate the behaviors of an AI diagnosis model under adversarial images generated by Generative Adversarial Network (GAN) models and to evaluate the effects on human experts when visually identifying potential adversarial images. Our GAN model makes intentional modifications to the diagnosis-sensitive contents of mammogram images in deep learning-based computer-aided diagnosis (CAD) of breast cancer. In our experiments the adversarial samples fool the AI-CAD model to output a wrong diagnosis on 69.1% of the cases that are initially correctly classified by the AI-CAD model. Five breast imaging radiologists visually identify 29%-71% of the adversarial samples. Our study suggests an imperative need for continuing research on medical AI model’s safety issues and for developing potential defensive solutions against adversarial attacks.

Suggested Citation

  • Qianwei Zhou & Margarita Zuley & Yuan Guo & Lu Yang & Bronwyn Nair & Adrienne Vargo & Suzanne Ghannam & Dooman Arefan & Shandong Wu, 2021. "A machine and human reader study on AI diagnosis model safety under attacks of adversarial images," Nature Communications, Nature, vol. 12(1), pages 1-11, December.
  • Handle: RePEc:nat:natcom:v:12:y:2021:i:1:d:10.1038_s41467-021-27577-x
    DOI: 10.1038/s41467-021-27577-x
    as

    Download full text from publisher

    File URL: https://www.nature.com/articles/s41467-021-27577-x
    File Function: Abstract
    Download Restriction: no

    File URL: https://libkey.io/10.1038/s41467-021-27577-x?utm_source=ideas
    LibKey link: if access is restricted and if your library uses this service, LibKey will redirect you to where you can use your library subscription to access this item
    ---><---

    References listed on IDEAS

    as
    1. Scott Mayer McKinney & Marcin Sieniek & Varun Godbole & Jonathan Godwin & Natasha Antropova & Hutan Ashrafian & Trevor Back & Mary Chesus & Greg S. Corrado & Ara Darzi & Mozziyar Etemadi & Florencia G, 2020. "International evaluation of an AI system for breast cancer screening," Nature, Nature, vol. 577(7788), pages 89-94, January.
    2. Scott Mayer McKinney & Marcin Sieniek & Varun Godbole & Jonathan Godwin & Natasha Antropova & Hutan Ashrafian & Trevor Back & Mary Chesus & Greg S. Corrado & Ara Darzi & Mozziyar Etemadi & Florencia G, 2020. "Addendum: International evaluation of an AI system for breast cancer screening," Nature, Nature, vol. 586(7829), pages 19-19, October.
    Full references (including those not matched with items on IDEAS)

    Most related items

    These are the items that most often cite the same works as this one and are cited by the same works as this one.
    1. Alexander P. L. Martindale & Carrie D. Llewellyn & Richard O. Visser & Benjamin Ng & Victoria Ngai & Aditya U. Kale & Lavinia Ferrante Ruffano & Robert M. Golub & Gary S. Collins & David Moher & Melis, 2024. "Concordance of randomised controlled trials for artificial intelligence interventions with the CONSORT-AI reporting guidelines," Nature Communications, Nature, vol. 15(1), pages 1-11, December.
    2. Joachim Meyer, 2024. "Doing AI: Algorithmic decision support as a human activity," Papers 2402.14674, arXiv.org, revised Apr 2024.
    3. Babak Abedin & Christian Meske & Iris Junglas & Fethi Rabhi & Hamid R. Motahari-Nezhad, 2022. "Designing and Managing Human-AI Interactions," Information Systems Frontiers, Springer, vol. 24(3), pages 691-697, June.
    4. Yuming Jiang & Zhicheng Zhang & Wei Wang & Weicai Huang & Chuanli Chen & Sujuan Xi & M. Usman Ahmad & Yulan Ren & Shengtian Sang & Jingjing Xie & Jen-Yeu Wang & Wenjun Xiong & Tuanjie Li & Zhen Han & , 2023. "Biology-guided deep learning predicts prognosis and cancer immunotherapy response," Nature Communications, Nature, vol. 14(1), pages 1-16, December.
    5. Shu Jiang & Jiguo Cao & Bernard Rosner & Graham A. Colditz, 2023. "Supervised two‐dimensional functional principal component analysis with time‐to‐event outcomes and mammogram imaging data," Biometrics, The International Biometric Society, vol. 79(2), pages 1359-1369, June.
    6. Helen M. L. Frazer & Carlos A. Peña-Solorzano & Chun Fung Kwok & Michael S. Elliott & Yuanhong Chen & Chong Wang & Jocelyn F. Lippey & John L. Hopper & Peter Brotchie & Gustavo Carneiro & Davis J. McC, 2024. "Comparison of AI-integrated pathways with human-AI interaction in population mammographic screening for breast cancer," Nature Communications, Nature, vol. 15(1), pages 1-12, December.
    7. Minkyu Shin & Jin Kim & Bas van Opheusden & Thomas L. Griffiths, 2023. "Superhuman Artificial Intelligence Can Improve Human Decision Making by Increasing Novelty," Papers 2303.07462, arXiv.org, revised Apr 2023.
    8. Sebastian Schleidgen & Orsolya Friedrich & Selin Gerlek & Galia Assadi & Johanna Seifert, 2023. "The concept of “interaction” in debates on human–machine interaction," Palgrave Communications, Palgrave Macmillan, vol. 10(1), pages 1-13, December.
    9. Armando Vargas-Palacios & Nisha Sharma & Gurdeep S. Sagoo, 2023. "Cost-effectiveness requirements for implementing artificial intelligence technology in the Women’s UK Breast Cancer Screening service," Nature Communications, Nature, vol. 14(1), pages 1-11, December.
    10. Juexiao Zhou & Haoyang Li & Xingyu Liao & Bin Zhang & Wenjia He & Zhongxiao Li & Longxi Zhou & Xin Gao, 2023. "A unified method to revoke the private data of patients in intelligent healthcare with audit to forget," Nature Communications, Nature, vol. 14(1), pages 1-13, December.
    11. Taynara de Oliveira Castellões & Paloma Maria Silva Rocha Rizol & Luiz Fernando Costa Nascimento, 2024. "Association between Premature Birth and Air Pollutants Using Fuzzy and Adaptive Neuro-Fuzzy Inference System (ANFIS) Techniques," Mathematics, MDPI, vol. 12(18), pages 1-12, September.
    12. Mélanie Roschewitz & Galvin Khara & Joe Yearsley & Nisha Sharma & Jonathan J. James & Éva Ambrózay & Adam Heroux & Peter Kecskemethy & Tobias Rijken & Ben Glocker, 2023. "Automatic correction of performance drift under acquisition shift in medical image classification," Nature Communications, Nature, vol. 14(1), pages 1-10, December.

    More about this item

    Statistics

    Access and download statistics

    Corrections

    All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:nat:natcom:v:12:y:2021:i:1:d:10.1038_s41467-021-27577-x. See general information about how to correct material in RePEc.

    If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.

    If CitEc recognized a bibliographic reference but did not link an item in RePEc to it, you can help with this form .

    If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.

    For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: Sonal Shukla or Springer Nature Abstracting and Indexing (email available below). General contact details of provider: http://www.nature.com .

    Please note that corrections may take a couple of weeks to filter through the various RePEc services.

    IDEAS is a RePEc service. RePEc uses bibliographic data supplied by the respective publishers.