01823nas a2200289   4500000000100000008004100001260003100042653001400073653002800087653001700115653000900132653003600141653002100177653001100198653001500209100001200224700001200236700001500248700001600263700002300279700002300302245013000325856007500455300001200530520096900542020002201511       2027                            d  c2027-01-01bAcademic Press10aAnimalgan10aArtificial intelligence10aData science10aDrug10agenerative adversarial networks10aMachine Learning10aToxGAN10aTransorgan1 aXi Chen1 aTing Li1 aWeida Tong1 aMinjun Chen1 aCatharine E. Krebs1 aGail A. Van Norman00aChapter 7 - The application of generative adversarial networks (GANs) as new approach methods (NAMs) toward nonanimal testing  uhttps://www.sciencedirect.com/science/article/pii/B9780443265143000199  a105-1193 aAnimal models have historically played a critical role in biomedical research and regulatory submissions. However, there is a growing movement toward nonanimal approaches driven by ethical concerns, cost-effectiveness, and limitations in predicting human outcomes. This chapter explores the transformative potential of artificial intelligence (AI) applications, particularly generative adversarial networks (GANs), in advancing nonanimal testing methodologies. It discusses how GANs can leverage existing data from traditional animal models, offering an alternative to animal studies as future avenues for integrating AI into the risk and safety assessment of drugs and chemicals. The chapter elucidates the principles of GANs with examples in toxicology through three case studies: ToxGAN, AnimalGAN, and TransOrGAN. In addition, it discusses the existing challenges and potential trajectories for leveraging this technology to enhance preclinical safety studies.  a978-0-443-26514-3