Artificial intelligence-enabled machine olfaction for volatile profiling and food quality decision-making

Monitoring food quality and aroma evolution remains a major challenge in food chemistry due to the complex dynamics of volatile organic compounds (VOCs). Although machine olfaction and artificial intelligence (AI) have advanced rapidly, their integration for food quality assessment remains limited. This review summarizes AI-enabled machine olfaction for food quality evaluation, focusing on VOC formation and transformation in food matrices. Relationships between biochemical degradation, VOC profiles, and olfactory signals are discussed. Key sensing technologies, including metal oxide semiconductor sensors, electrochemical sensors, bio-inspired materials, and electronic nose (e-nose) systems, are reviewed in terms of analytical performance and operational trade-offs relative to reference VOC instruments. Machine learning (ML) and deep learning (DL) methods for odor pattern recognition and multimodal data fusion are also highlighted. AI-enabled machine olfaction enables conversion of complex VOC signals into quantitative indicators for freshness evaluation, spoilage detection, authenticity assessment, and cold-chain monitoring. Key limitations include sensor drift, environmental variability, dataset scarcity, and weak model transferability. Future work should focus on chemically informed sensing strategies and explainable AI to improve robustness and regulatory acceptance.