Article Text
PDFCommentary
Trustworthy medical AI systems need to know when they don’t know
Statistics from Altmetric.com
Footnotes
Funding TG is supported by the Deutsche Forschungsgemeinschaft (BE5601/4-1; Cluster of Excellence ‘Machine Learning—New Perspectives for Science’, EXC 2064, project number 390727645).
Competing interests None declared.
Provenance and peer review Commissioned; internally peer reviewed.
↵Note, while Durán and Jongma emphasise that physicians should not blindly defer to algorithmic decisions due to choices of value, my account is confined to the epistemic part of his paper.
Linked Articles
Read the full text or download the PDF:
Other content recommended for you
- Randomised controlled trials in medical AI: ethical considerations
- Who is afraid of black box algorithms? On the epistemological and ethical basis of trust in medical AI
- Public perceptions on the application of artificial intelligence in healthcare: a qualitative meta-synthesis
- Design publicity of black box algorithms: a support to the epistemic and ethical justifications of medical AI systems
- Limits of trust in medical AI
- Artificial intelligence for diabetic retinopathy in low-income and middle-income countries: a scoping review
- AI support for ethical decision-making around resuscitation: proceed with care
- Evaluation framework to guide implementation of AI systems into healthcare settings
- Development and validation of a deep learning system to screen vision-threatening conditions in high myopia using optical coherence tomography images
- Does “AI” stand for augmenting inequality in the era of covid-19 healthcare?