Entrectinib, a potent, CNS-active, TRK and ROS1 inhibitor is associated with occurrence of bone fractures, particularly in pediatric patients. We embarked on computer-guided and experimental investigations to identify the underlying mechanism that should help safe dosing of children receiving this drug. An artificial intelligence (AI)-based approach was used to allow for hypothesis generation regarding drug effects in a clinical setting and mechanistic insights on safety and efficacy. This was followed by confirmatory in vitro studies using human bone cell cultures. The AI-based hypothesis suggested an effect via inhibition of TRK and JAK2, as molecular targets of Entrectinib, underlying the fractures. Subsequent experiments using human bone cells showed that Entrectinib reduced markers of osteoblastic- and mineralization-activity. Conversely, Entrectinib had stimulating effects on osteoclast function indicative of bone resorption. 3D osteoblast-osteoclast cultures revealed reduced stiffness after Entrectinib treatment and this was pronounced in the absence of 17β-estradiol, characteristic for juvenile bone. Pathway analysis revealed inhibition of bone growth regulators TGF-β and BMP which is suggested to underlie the in silico-proposed JAK2 effect. Pre-treatment with Vitamin D led to partially restored osteoblast function in vitro, while osteoclast function was suppressed. These experiments suggest an on-target effect of Entrectinib underlying increased fractures particularly in children with developing bones, offering potential mitigation strategies if these observations can be confirmed with data from ongoing clinical trials.