J Am Coll Radiol. 2018 Mar;15(3 Pt B):497-498. doi: 10.1016/j.jacr.2018.01.029. Carlos RC1, Kahn CE2, Halabi S3.
A quick search on PubMed.gov for each of the terms that have been used to describe some aspect of data science exceeded 10,000 results per term with artificial intelligence (AI) returning 74,250 results. Moehler introduces a future with a need for fewer radiologists with robots hunched at the view box [1]. Although the death of radiology due to machine learning and AI has been greatly exaggerated, the field of imaging and the nature of the imaging care will change. Our cracked crystal ball cannot opine on how fast the change will happen, whether it will be the frog gradually boiling or Wile E. Coyote falling as the road abruptly runs out. Nor can it predict the substance of the change; will we be plugged into the machine or will the machine be plugged into us? Nevertheless, despite the dystopian futures we have binge-watched over the last year, we will make a positive prediction—every radiologist who can adapt to the changes ahead will have a job. The rub is identifying what these changes might be and what adaptation might look like—from a radiologist’s perspective, an enterprise’s perspective, and society’s perspective. We have invited a diverse group of forward thinkers to speculate on the near future to provide some trail markers that we can follow, knowing that the trail itself remains to be fully mapped.