TY  - JOUR
T1  - Building Trust in the Application of Machine Learning Algorithms for Rare Disease Diagnosis
A1  - Lukas Lembo
A1  - Marcello Barra
A1  - Alexander Iriti
JF  - Asian Journal of Ethics in Health and Medicine
JO  - Asian J Ethics Health Med
SN  - 3108-5059
Y1  - 2023
VL  - 3
IS  - 1
DO  - 10.51847/Mo7NXmiBnA
SP  - 26
EP  - 39
N2  - As AI becomes increasingly integrated into healthcare and computerised systems influence clinical decision-making, addressing both trust in and the trustworthiness of AI tools is critical. Focusing on computational phenotyping (CP) for diagnosing rare diseases in dysmorphology, this paper investigates the conditions under which medical AI tools employing machine learning can be trusted. Semi-structured qualitative interviews (n = 20) were conducted with stakeholders involved in designing or using CP systems, including clinical geneticists, data scientists, bioinformaticians, industry representatives, and patient support group spokespersons. Interview data were analysed using the method of constant comparison. Participants highlighted the centrality of trust in the deployment of CP technology for rare disease diagnosis. Trust was conceptualized in two interconnected ways. First, they emphasized the importance of trust relationships: patients must trust the clinicians using AI tools, and clinicians must trust AI developers, in order to facilitate adoption. Second, participants stressed the need for trust in the technology itself, or epistemic trust in the knowledge it generates. CP tools may be viewed as more trustworthy if their reliability and accuracy are verifiable and if the users or developers themselves are trusted. The findings indicate the necessity of intentionally designing AI systems that are reliable and confidence-worthy for healthcare applications. Moreover, establishing robust processes—such as randomized controlled trials (RCTs) or frameworks ensuring accountability, transparency, and responsibility—can help affirm the epistemic trustworthiness of these tools.
UR  - https://smerpub.com/article/building-trust-in-the-application-of-machine-learning-algorithms-for-rare-disease-diagnosis-fjmvhd2839doshn
ER  -