The growing demand for lightweight, high-performance, and fire-safe polymer materials has accelerated research into advanced flame-retardant composites. Traditional experimental approaches to designing sustainable flame-retardant biodegradable polymer composites still rely heavily on empirical formulation and iterative testing, which are time-consuming and costly, and they often struggle to capture the coupled effects of chemical composition, processing conditions, and material performance. Recent advances in artificial intelligence (AI) provide opportunities to address these challenges by learning formulation–structure–performance relationships from curated datasets and by translating materials chemistry and flame-retardant mechanisms into data-ready descriptors and targets. This review summarizes recent progress of AI-assisted approaches to design sustainable flame-retardant biodegradable polymer composites, emphasizing machine learning, deep learning, and active learning methods for predicting and optimizing key fire performance metrics, including limiting oxygen index and heat release-related parameters. Biodegradable-specific limitations, including narrow processing window, thermal degradation, and moisture sensitivity, are discussed in the content of descriptor selection and constraint-aware optimization, together with the role of interpretable/explainable models in supporting experimentally actionable guidance. Current challenges such as limited data availability, protocol variability, model transferability, and interpretability are highlighted, and emerging solutions, including data harmonization, standardized fire testing, and physics-informed models are outlined. AI-assisted strategies are expected to play a central role in accelerating efficient, sustainable, halogen-free, and performance-driven development of next-generation flame-retardant biodegradable polymer composites.
Zhang et al. (2026) studied this question.