The article deals with adverse drug reactions (ADRs) and emphasizes the need for their systematic characterization, including delayed or atypical reactions.[5] He cautions that clinically significant adverse effects associated with the immune system with multiple organ involvement may occur long after the initiation of treatment or even after discontinuation of the drug, which may lead to repeated exposure and life-threatening conditions if the cause is not detected early.[5] The article emphasizes that models based on machine learning have a positive impact on proactive risk identification and stratification.[5] An example is the development of a logistic regression-based scoring tool to estimate the risk of adverse events associated with semaglutide, which identified gastrointestinal comorbidities, alcohol consumption, and concomitant use of other medications as key risk factors.[5] The article also points to molecular interactions between drugs and biological systems that lead to diverse adverse effects, such as inhibition of GPR83-AKT signaling by ibrutinib, which induces hair apoptosis and hearing loss.[5]