A critical review of artificial intelligence in biomaterials science from machine learning and explainable AI to...

Artificial intelligence (AI), especially machine learning (ML) and deep learning (DL), is rapidly changing biomaterials science. These computational tools enable researchers to analyze large and complex datasets, discover new materials, and improve the design and performance of existing ones. ML and DL algorithms learn from experimental and simulation data to predict material properties, guide synthesis, and optimize manufacturing conditions, often faster than conventional trial-and-error approaches. In recent years, AI-based methods have been applied across biomaterials research, including predicting the biocompatibility and mechanical strength of new materials, improving additive manufacturing (AM) processes, and enhancing the precision of biofabrication and tissue engineering. AI tools are increasingly applied to real-time quality assessment and adaptive control during material production, enabling the design of "smart" biomaterials that respond dynamically to environmental and biological signals, with potential uses in regenerative medicine, healthcare devices, and sustainable materials. This emerging field still faces key challenges, such as the difficulty of obtaining large and reliable datasets needed to train accurate AI models. Another critical issue is the interpretability of complex ML and DL models. Understanding why an algorithm makes a certain prediction is essential for building trust and guiding experimental validation. To address these issues, researchers are turning to explainable AI (XAI) approaches that provide greater transparency and insight into model behavior. This review summarizes recent progress in applying ML, DL, and XAI to biomaterials science. It also highlights the main opportunities and challenges for developing intelligent, real-time adaptive materials for healthcare, regenerative medicine, and sustainable material design.