ARPA‑H awarded up to $21 million to an industrial‑academic consortium led by Inductive Bio to build next‑generation, human‑biology‑driven AI models for drug toxicity prediction. Partners include Amgen, Baylor College of Medicine, Cincinnati Children’s, Torch Bio and others; the DATAMAP project will generate data from organoids, ex vivo human tissues and microphysiological systems to train AI models for drug‑induced liver injury and cardiotoxicity. The team aims to reduce reliance on animal testing and improve preclinical safety prediction by combining advanced human model systems with mechanistic AI. Inductive’s Ben Birnbaum said animal tests often fail to capture human responses and the project will work with the FDA to validate regulatory applications. If successful, the platform could cut late‑stage attrition driven by unpredicted toxicity and change how sponsors prioritize candidates, while offering regulators new validation data for human‑centric safety assessment workflows.