Abstract
Radiology artificial intelligence (AI) is advancing however its adoption faces fragile foundations that threaten sustainability. Despite bold promises of efficiency and accuracy, current deployment is undermined by weaknesses in economics, evidence, infrastructure, human factors, regulation, security, and environmental impact. Nearly 90% of radiology AI studies report process metrics rather than patient outcomes, while hidden costs elevate ownership to 400% to 500% of subscription fees. Technical fragilities include 25% or greater performance loss with routine protocol or scanner shifts, compounded by vendor consolidation that has eliminated 63% of companies since 2020, creating migration costs averaging 180,000 dollars per exit. Human factor challenges, including automation bias and progressive deskilling, intersect with regulatory requirements that mandate continuous evidence generation. Security risks and environmental costs remain underrecognized. This review introduces frameworks including risk assessment matrices, compliance guides, procurement checklists, evidence standards, lifecycle calculators, and implementation protocols to enable sustainable, patient centered, value driven integration.
Article Type
Review
First Page
195
Last Page
218
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Recommended Citation
Alhasan, Mustafa S.
(2026)
"Artificial Intelligence in Radiology: Hidden Fragilities and the Path to Resilience,"
Saudi Medical Journal: Vol. 47:
Iss.
2, Article 3.