STAT+: Radiology captures the tension between AI’s potential and concern for patients
This year’s meeting of the Radiological Society of North America surfaced the tension between AI's potential and clinicians' emphasis on improving patient outcomes.
Radiology has long led the way in the application of artificial intelligence in medicine. More than three-quarters of AI and machine learning devices authorized by the Food and Drug Administration work with medical images — and that dominance was evident at this year’s meeting of the Radiological Society of North America, where AI vendors gathered to tout their latest offerings.
“How clinical AI is changing radiology for good,” Philips promised to share in a sponsored session. Nuance, the AI-driven voice recognition company acquired by Microsoft last year for $16 billion, sponsored a lunch session on the promise of generative AI to “raise the bar on quality and efficiency,” echoing the frequent claim that AI can solve the plague of physician burnout and staff shortages.
But as pitches for the next generation of AI dominated the demonstration floor, there was tension between excitement and skepticism of the technology. The meeting’s scientific sessions reflected a more mixed reality — one marked with uncertainty and some concern about the real impact of AI tools on patient outcomes.
What's Your Reaction?
