The synergy of artificial intelligence and organ-on-a-chip: building a next-generation intelligent engine for drug...

Drug development has long been constrained by high costs, lengthy timelines, and low success rates, primarily due to the insufficient human physiological fidelity of traditional preclinical models. Artificial intelligence (AI) has accelerated computational drug discovery, but its models inherit training-data bias and are rarely validated against human-relevant experiments. Organ-on-a-chip (OoC) systems produce human-relevant, multimodal physiological data, yet the sheer volume of data they generate overwhelms conventional processing pipelines. With regulators now encouraging non-animal testing, integrating AI with OoC has become a practical route forward. This review examines how the two technologies reinforce each other across the OoC lifecycle, surveys progress toward clinical and industrial translation, and considers longer-term paradigms such as digital twins and programmable virtual humans. Major hurdles remain in hardware performance, biological fidelity, data standardization, and model interpretability. Closer collaboration across sectors would help this combined dry-wet approach reduce animal testing and make drug research more predictive and more human-relevant.