Artificial intelligence, particularly deep learning (DL), is poised to transform the field of cardiovascular imaging. While still a relatively young discipline, there has been an explosion of research in DL for cardiovascular imaging as investigators seek to build systems designed to segment cardiac chambers, automate functional assessment, detect disease states, and predict prognosis—all using raw imaging data.1-3 Innovative systems have been designed to guide a novice to obtain diagnostic bedside cardiac ultrasonography4 or noninvasively estimate fractional flow reserve on coronary computed tomography angiography (CCTA).5 Such is the promise of DL, a powerful technology that requires little input from humans in the training/development process and often produces highly accurate results. However, this human input is vital for ensuring valid, unbiased, and generalizable model outputs.