Breast imaging AI falls flat: only 9% of radiologists see fewer biopsies
Curated by the Inblix editorial team
The gap between AI hype and clinical reality just got measured, and it’s not flattering. A survey of 215 Society of Breast Imaging members, published in Clinical Imaging, found that roughly half already use FDA-approved AI tools for breast cancer detection, with another 11 percent planning to adopt them. But the enthusiasm stops there. Lead author Joud Almogati at UC San Diego Health reports that only a few radiologists consider AI a deciding factor in their work.
The numbers tell a story of dashed expectations. Only 35 percent of users report lower recall rates, despite 59 percent expecting that benefit before adoption. Even starker: just 9 percent see fewer unnecessary biopsies, against 36 percent who hoped AI would reduce them. Burnout relief is similarly underwhelming — 29 percent report feeling less burned out, while 56 percent expected meaningful relief. Most radiologists now view AI as a second opinion rather than a transformative tool. Costs and lack of institutional support remain the biggest barriers to broader adoption.
This shouldn’t surprise anyone who’s watched AI predictions cycle through the hype machine. About a decade ago, prominent AI researchers confidently predicted radiologists would soon lose their jobs to algorithms. That didn’t happen. Similar claims are making the rounds again, this time targeting knowledge work on computers. Nvidia CEO Jensen Huang has a name for this kind of forecasting: a “God complex” among the prophets of AI-driven job loss.
The irony is that breast imaging is one of the most mature application areas for medical AI, with multiple FDA-cleared tools on the market. If the gap between expectation and reality is this wide here, it’s a warning sign for newer, less validated AI applications in healthcare. The technology works, radiologists are using it, but the promised transformation remains stubbornly out of reach. The question now is whether vendors will recalibrate their marketing to match what the tools actually deliver — or keep selling the fantasy.
💡 Key Takeaways
- About half of surveyed breast imaging radiologists already use FDA-approved AI tools, but only a few consider AI a deciding factor in their work.
- Only 9 percent of radiologists report fewer unnecessary biopsies from AI, versus 36 percent who expected that benefit — the largest expectation gap in the survey.
- The gap mirrors a decade-old pattern: AI researchers predicted radiologists would lose their jobs, and that never materialized.
- Cost and lack of institutional support, not skepticism about the technology itself, are the biggest barriers to broader adoption.
Keep reading: See related articles below for more coverage on this topic.
Get smarter about AI
The sharpest AI news, curated daily. Delivered free to your inbox.