The greatest health care and pharmaceutical problems remain adverse drug reactions and chemical toxicities. Because of the extensive use of animal testing, conventional toxicology approaches often do not have the mechanistic insights required to relate molecular perturbations to their adverse biological outcomes. To address this limitation, adverse outcome pathway architecture is a sequence of causally related events beginning with a molecular initiating event, followed by key events, and concluding with an adverse outcome. This study describes the construction and applications of adverse outcome pathways in chemical risk assessment, predictive toxicology, and regulatory decision-making. More recent developments, including quantitative adverse outcome pathways and adverse outcome pathway networks, have enhanced the ability to model biological complexity and dose-response interactions. Adverse outcome pathways provide a solid mechanistic basis for toxicity prediction and allow the transition of modern toxicology to non-animal, human-relevant testing systems, although barriers such as a lack of quantitative data and regulatory acceptability stand.