Letter: Spillover Effects in Randomized Evaluations of Translational AI
This letter expands on Leuchter et al.’s “two‑way moving target” in assessing clinical AI, adding that spillover from limited resources can bias trial results and should be accounted for in real‑world evaluations.