Imagine developing a life-saving medication without knowing how it will affect a living body. To address this, scientists have relied on animal models—a practice dating back over 2,400 years that remains a cornerstone of modern biomedical research. Researchers commonly use mice and rats to study diseases and evaluate the safety and efficacy of drugs before trials in humans. However, in April 2025, the Food and Drug Administration (FDA) published a roadmap outlining plans to reduce unnecessary animal testing by advancing alternative research methods. Considering both the strengths and weaknesses of animal models, is reducing their use necessary?

Animal models continue to play an important role in biomedical research because they have demonstrated historical value in advancing treatments, promoting safety, and helping researchers understand disease progression. At the same time, scientific limitations, ethical concerns, and high failure rates in human clinical trials have encouraged researchers to develop New Approach Methodologies (NAMs) as potential alternatives.

Animal models have demonstrated remarkable historical success in the biomedical field. Research involving mice has contributed to 42 Nobel Prize-winning discoveries, with rats contributing to 31 more (Foundation for Biomedical Research, 2020). Animal models played an important role in discovering insulin for Type I diabetes, developing treatments for hypertension, cancer immunotherapy, and anti-inflammatory therapies (Guo et al., 2024). When we look at some of the most significant advances in modern medicine, animals were central to research behind the scenes.

Another strength of animal models is that many animals, especially mammals, share similar anatomical and physiological systems with humans. Both rely on the same organ systems and cellular structures, allowing researchers to study disease development and evaluate potential treatments before testing them in people (Sinoussi & Montagutelli, 2015). Safety is another reason animal testing remains part of the research process. Before treatments reach human volunteers, researchers must identify potential risks involving toxicity, dosage, and early signs of effectiveness. The COVID-19 vaccine demonstrates this process well, as researchers used mice, rats, and monkeys during preclinical testing to confirm its safety and effectiveness before human trials.

Despite these strengths, animal models are far from perfect. One of the greatest challenges is that animal experiments often fail to accurately predict human responses. Although mice and humans share over 95% homologous genes, important differences still exist in gene expression, immune responses, cellular receptors, and overall physiology (Sinoussi & Montagutelli, 2015). Scientists also frequently induce diseases artificially in animals, making it difficult to replicate the full complexity of naturally occurring human conditions. Consequently, treatments that appear successful in animals often fail during human clinical trials. In fact, approximately 86% of investigational drugs fail during clinical testing because animal models cannot fully predict how humans will respond (Kwon, 2026).

Animal testing is also expensive, time-consuming, and ethically controversial. According to the National Library of Medicine, rodent testing in cancer therapeutics can cost between two and four million dollars while requiring up to five years to complete (Norman, 2019). Beyond financial costs, millions of animals are used in research each year, raising ethical concerns about balancing scientific advancement with animal welfare (Kiani et al., 2022). These scientific and ethical limitations have motivated researchers to develop alternative approaches that reduce reliance on animal testing while improving predictions of human biology.

These alternatives, known as New Approach Methodologies (NAMs), include organs-on-chips, 3D organoids, and AI-driven computational models. Rather than replacing every aspect of animal research immediately, these technologies aim to complement and eventually reduce the use of animal models by providing more accurate, human-specific data during preclinical testing.

One of the most promising developments in New Approach Methodologies (NAMs) is the use of organs-on-chips and 3D organoids. These technologies rely on human-derived induced pluripotent stem cells (iPSCs), which can be differentiated into various tissue types while maintaining human-like genetic characteristics. Organoids are able to model human diseases while mimicking human development and cellular diversity, making them highly valuable for personalized medicine and studying genetic disorders. Similarly, organs-on-chips use fluid-filled channels to expose human cells to drugs or disease-related conditions while recreating important biological processes (Kwon, 2026). Integrating organs-on-chips with organoids creates continuous fluid flow that mimics human blood circulation and mechanical functions (Papamichail, 2025). Compared to animal models, these systems can better reflect human-specific responses to drugs because they use human cells and recreate biological environments that cannot be fully replicated in animals. As these technologies continue to advance, they have the potential to significantly reduce unnecessary animal testing.

Computational models and generative artificial intelligence also provide promising alternatives through their ability to predict toxicological outcomes. These methodologies analyze large datasets, including chemical properties, laboratory test results, toxicity records, and human-derived biological data, to accurately predict in vivo responses (Kwon, 2026). Machine learning programs can identify complex reactions such as skin sensitization and liver injury before human trials begin. For example, researchers developed an in silico skin-sensitization model using data from approximately 430 human chemicals, allowing the system to accurately identify substances likely to trigger allergic skin reactions (Kwon, 2026). Additionally, generative AI can accelerate drug discovery by combining biological and chemical data to detect toxins and identify potential drug candidates in a fraction of the time required for traditional animal research. Together, these computational approaches offer researchers faster, more efficient, and increasingly accurate methods for evaluating the safety of new therapies.

Despite these advances, current NAMs still have technical limitations in fully replicating the complexity of human physiology. Because these systems often focus on individual tissues or specific biological mechanisms, they struggle to reproduce whole-body interactions between multiple organs. Processes such as tissue aging, hormonal regulation, immune system interactions, and long-term disease progression remain difficult to model because they depend on coordinated communication throughout the body (Kwon, 2026). Although organoids, organs-on-chips, and computational models are highly effective for studying specific mechanisms, they do not yet capture the full complexity of human biological systems.

Overall, animal models have made extraordinary contributions to biomedical research by advancing medical discoveries, improving patient safety, and deepening our understanding of human disease. However, their scientific limitations, high clinical failure rates, financial costs, and ethical concerns demonstrate why alternative methodologies are becoming increasingly important. New Approach Methodologies represent a major step toward safer, more ethical, and more human-relevant biomedical research. While current technologies cannot yet replace animal models entirely, continued advances in human cell models, organs-on-chips, and artificial intelligence have the potential to greatly reduce reliance on animal testing and eventually transform the future of preclinical research.

References:

Barre-Sinoussi, F., & Montagutelli, X. (2015). Checking your Browser – reCAPTCHA. Animal Models are Essential to Biological Research: Issues and Perspectives, https://pmc.ncbi.nlm.nih.gov/articles/PMC5137861/

Commissioner O. (2026, April 20). FDA Achieves Year 1 Goals in Reducing Animal Testing in Drug Development. U.S. Food and Drug Administration, https://fda.gov/news-events/press-announcement/fda-achieves-year-1-goals-reducing-animal-testing-drug-development

Guo, H., Xu, X., Zhang, J., Du, Y., Yang, X., He, Z., Zhao, L., Liang, T., & Guo, L. (2024). The Pivotal Role of Preclinical Animal Models in Anti-Cancer Drug Discovery and Personalized Cancer Therapy Strategies. Pharmaceuticals, 17(8), 1048. https://doi.org/10.3390/ph17081048

Kiani, A.K. (2022). Checking your Browser – reCAPTCHA. Publication_Title, https://pmc.ncbi.nlm.nih.gov/articles/PMC9710398/

Kwon, D. (2026). The age of animal experiments is waning. Where will science go next?. Nature. https://www.nature.com/articles/d41586-026-00563-3

Normal, G. (2019). Limitations of Animal Studies for Predicting Toxicity in Clinical Trials. JACC: Basic to Translational Science, 4(7), 845-854., https://pubmed.ncbi.nlm.nih.gov/31998852/

Papamichail, L. (2025, March 11). Checking your browser – reCAPTCHA. National Library of Medicine. https://pmc.ncbi.nlm.nih.gov/articles/PMC11933005/

2020 Nobel Prize in Medicine Awarded for Hepatitis C Discovery. (2025). 2025 Nobel Prize in Medicine: Immune Discoveries With Mice. Publication_Title, https://fbresearch.org/2025-nobel-prize-mice