AI tools for breast cancer detection are not meeting the high expectations of radiologists, according to a recent survey. The study, published in Clinical Imaging, surveyed 215 members of the Society of Breast Imaging. It found that about half of the participants already use FDA-approved AI tools, and 11 percent plan to. However, the tools are not seen as a major factor in decision-making, with only a few considering them a deciding factor. The tools help, but not as much as radiologists had hoped. Source: thedecoder
Only 35 percent of radiologists report lower recall rates with AI tools, far below the 59 percent who expected this improvement. Similarly, 9 percent see fewer unnecessary biopsies, compared to the 36 percent who anticipated this outcome. Additionally, 29 percent report less burnout, below the 56 percent who expected relief. Most radiologists view AI mainly as a second opinion rather than a primary diagnostic tool. Cost and lack of institutional support are the main barriers to wider adoption. Source: thedecoder
The AI industry may have set itself up for this gap, as prominent researchers predicted about ten years ago that radiologists would soon lose their jobs. Similar claims are resurfacing, this time about knowledge work on computers. Nvidia CEO Jensen Huang has criticized such predictions as a 'God complex' among AI proponents. Source: thedecoder
