Artificial intelligence (AI), machine learning and smart sensing technologies are emerging as tools to shift mycotoxin management from reactive testing to predictive and real-time risk surveillance, according to a 2026 review published in the Journal of Stored Products Research.
The review analysed 65 studies from 1,200 records covering advances in AI-driven predictive modelling and smart surveillance for mycotoxin risk management across global food supply chains.
From Detection to Prediction
Conventional mycotoxin testing relies largely on laboratory-based methods such as HPLC, ELISA and mass spectrometry. While these techniques offer high accuracy, they can involve significant costs, specialised infrastructure and delays between sampling and results.
AI-based systems offer a different approach by combining machine learning with sensor and environmental data to forecast contamination risks before they become critical. Models can analyse variables such as temperature, humidity, moisture, crop conditions, storage parameters and historical contamination data to identify patterns associated with fungal growth and toxin production.
Smart Surveillance Across the Supply Chain
The review highlights the potential of integrating AI with the Internet of Things (IoT), smart sensors, remote sensing, cloud computing and edge computing.
IoT-enabled sensors can continuously monitor conditions in grain stores and other facilities, while AI models can analyse the data and generate early-warning alerts. Remote sensing technologies, including satellite and drone-based systems, could further help identify environmental conditions associated with mycotoxin risk.
For feed manufacturers, such systems could eventually support risk-based raw-material screening, storage management and targeted laboratory testing, allowing resources to be directed towards higher-risk consignments.
Climate Change Adds Complexity
Changing temperature and rainfall patterns are also altering fungal growth and mycotoxin contamination risks. The review points to the importance of incorporating climate and geospatial data into predictive models to identify potential contamination hotspots and improve early-warning capabilities.
Technology Still Needs Validation
Despite the promise, AI-based mycotoxin surveillance remains largely at the pilot and experimental stage. The researchers identify several barriers, including limited standardised datasets, weak model transferability between regions, fragmented data from different parts of the supply chain and limited integration of climate, storage and transport information.
The review also highlights the need for explainable AI, interoperable data systems, real-time IoT monitoring and global data-sharing platforms. Blockchain-based traceability and hybrid AI systems could further strengthen transparency and supply-chain resilience.
For the feed industry, the emerging technology could eventually move mycotoxin control beyond periodic laboratory testing towards a continuous, predictive and risk-based system—helping feed mills identify problems earlier, optimise testing and strengthen raw-material quality management.
Source: Journal of Stored Products Research, 2026
