Year: 2025
Type: article
Abstract: Animal models have long underpinned biomedical and pharmaceutical research; however, their use is constrained by species-specific biological differences, ethical concerns, high costs, and limited translational relevance to humans. Recent advances in artificial intelligence (AI) and in silico modelling provide increasingly human-relevant alternatives capable of simulating complex physiological and pathological processes, offering promising strategies to reduce reliance on animal experimentation while improving preclinical precision. This narrative review examines the potential of AI-driven in silico simulations as replacements or complementary tools to traditional animal models, evaluates their strengths and limitations, and assesses their alignment with the global 3Rs (Replacement, Reduction, Refinement) principle. Literature published between 2015 and 2025 was reviewed, focusing on the limitations of animal models, emerging AI and computational approaches, applications in drug discovery and toxicity prediction, evolving regulatory perspectives, and challenges to widespread adoption, with additional insights drawn from the uploaded manuscript. The findings indicate that AI-based models often outperform animal studies in predicting toxicity, pharmacokinetics, and drug disease interactions, while approaches such as virtual screening, physiologically based pharmacokinetic modelling, digital organ twins, and in silico clinical trials substantially reduce development time and cost and enhance human relevance. Although regulatory acceptance is increasing, issues related to data quality, model validation, and standardization persist.
Authors: Ayush Verma, Alankar Shrivastava
Institution: Amity University Chhatisgarh
Language: English
Subfield: Small Animals
Field: Veterinary
Domain: Health Sciences
Open Access status: diamond
Funder: Amity University
OpenAlex ID: W7125797049
