The serendipity paradox of artificial intelligence in oncologyOA
Artificial intelligence(AI)in oncology is often designed to exploit known patterns,creating a“serendipity paradox”:reliance on supervised learning and average performance systematically filters out the rare,anomalous,and unclassifiable.This editorial dissects three mechanisms driving this loss of surprise and proposes actionable strategies:anomaly detection as a primary objective,uncertainty-aware human-AI interfaces,and noise-preserving data governance,to reorient AI toward discovery of the genuinely unknown.
Zejia Mao;Bo Xu
Chongqing Key Laboratory of Intelligent Oncology for Breast Cancer,Intelligent Oncology Innovation Center Designated by the Ministry of Education,Chongqing University Cancer Hospital,Chongqing University School of Medicine,Chongqing 400030,ChinaChongqing Key Laboratory of Intelligent Oncology for Breast Cancer,Intelligent Oncology Innovation Center Designated by the Ministry of Education,Chongqing University Cancer Hospital,Chongqing University School of Medicine,Chongqing 400030,China
医药卫生
average performanceactionable strategies anomaly detectionsupervised learninganomaly detectionartificial intelligence aiartificial intelligenceserendipity paradoxoncology
《Intelligent Oncology》 2026 (3)
P.1-3,3
评论