Another day, another (worthwhile) rumination on healthcare AI

Every week now seems to bring a flurry of think pieces on how artificial intelligence will change everything, including healthcare (and especially radiology). Few break any new developments or insights, but an article posted to The Conversation website Dec. 7 does an uncommonly fine job explaining how much harder it will be for AI to diagnose a disease than, say, recognize a face.

“Radiologists might disagree slightly when interpreting a scan where blurring and only very subtle features can be observed,” writes Olivier Salvado, PhD, a biomedical informaticist with Australia’s federal Commonwealth Scientific and Industrial Research Organisation (CSIRO). “Inferring a diagnosis from measurement with errors and when the disease is modulated by unknown genes often relies on implicit know-how and experience rather than explicit facts.”

Still, Salvado points out, large medical data sets are indeed being assembled.

“New technologies to overcome the lack of certainty are being developed. Novel ways to establish causation are emerging. … This area is moving fast and tremendous potential exists for improving efficiency and health.”

He points out various directions and techniques in play or coming soon.

Read the piece: