The exponential growth in worldwide smartphone adoption has led to mounting concerns about problematic smartphone use (PSU), with some studies of variably defined PSU yielding prevalence rates as high as 20 to 30%. The present study aims to (a) estimate the prevalence of PSU while emphasizing functional impairment, and (b) examine the degree to which a set of demographic, contextual, behavioral and mental health variables can predict PSU. A representative U.S. sample of 1,989 participants completed an online survey using a cross-sectional design aimed at examining PSU. Data were analyzed using two different but complementary machine learning algorithms. The results indicate a PSU prevalence of 0.75% to 1.2 % when a functional impairment criterion is required. The most important predictors of PSU are related to smartphone use behavior, followed by contextual factors (being a parent of under-age children, increase in use during the COVID-19 pandemic), age and some psychopathological factors. Overall, the study suggests that the prevalence of functionally impairing PSU may be much lower than what has been reported based on less stringent definitions. The results highlight the necessity to account for impact on key spheres of functioning in diagnosing PSU to avoid pathologizing a ubiquitous behavior that may be intensive but not necessarily pathological. They also open up relevant perspectives for the prevention of PSU.