A 2016 prediction by Nobel-winning AI pioneer Geoffrey Hinton that radiologists would be replaced by computers within five years has proven exaggerated, as the field's ranks are expected to grow by 26 percent or more over the next three decades. However, radiology remains medicine's hot spot for AI; as of early 2026, about three-quarters of the 1,400 AI-enabled medical devices cleared by the FDA were for radiology. Some tools improve efficiency by drafting reports or flagging urgent images, while others identify abnormalities invisible to the human eye. An analysis of 43 clinical trials found AI-assisted colonoscopies reveal more polyps than conventional ones. Improving accuracy is important because average human error rates involving diagnostic images range from 3 to 5 percent, translating to about 40 million errors worldwide each year. Yet the solution is not simply replacing humans with machines. Radiologist Curtis Langlotz notes that even if AI is more reliable on average, it still makes mistakes humans would not, putting radiologists in the role of evaluating each AI decision.