The four dimensions of contestable AI diagnostics - A patient-centric approach to explainable AI

作者：

Highlights：

• Many high-level reports endorse’ explainability’ as a necessary feature of good AI.

• From a patient-centric view explainability is best understood as effective contestability.

• Effective contestability of AI diagnostics requires information about data use, bias, accuracy, and the diagnostic role of AI.

摘要

•Many high-level reports endorse’ explainability’ as a necessary feature of good AI.•From a patient-centric view explainability is best understood as effective contestability.•Effective contestability of AI diagnostics requires information about data use, bias, accuracy, and the diagnostic role of AI.

论文关键词：AI diagnostics,Contestability,Explainability,Health data,Bias,Performance,Organisation of diagnostic labour

论文评审过程：Received 30 January 2020, Revised 30 April 2020, Accepted 3 June 2020, Available online 9 June 2020, Version of Record 15 June 2020.

论文官网地址：https://doi.org/10.1016/j.artmed.2020.101901