AI Reshapes Crop Breeding for Global Food Security

180

A new study published in the International Journal of Molecular Sciences highlights how artificial intelligence (AI) is transforming the breeding of disease-resistant crops. Titled “Artificial Intelligence-Assisted Breeding for Plant Disease Resistance,” the review details how AI technologies are revolutionising plant pathology, from real-time disease detection to predictive genomics.

With plant diseases causing severe crop losses globally, the study positions AI as essential to modern agriculture, offering tools that enhance detection accuracy, shorten breeding cycles, and forecast resistance traits across diverse crop types.

From Image Recognition to Decision Support

One of AI’s most immediate contributions is image-based disease detection. Models like convolutional neural networks (CNNs) and real-time tools such as YOLO (You Only Look Once) are now being deployed via drones and mobile devices, enabling fast and accurate diagnosis in the field. Through a review of over 340 papers from 2020 to 2025, the study identifies CNNs, YOLO, attention mechanisms, and vision transformers as leading technologies in this space.

Advanced hybrid systems like YOLO-GPT and CTDUNet combine visual recognition with large language models (LLMs) such as GPT-4, allowing not only detection but also automated guidance for farmers. Lightweight tools like MobileNet are helping bring AI to resource-limited environments.

Genomics, Multi-Omics, and New Frontiers

Beyond visual analysis, AI is powering the next wave of predictive breeding. Deep learning models now outperform traditional methods in identifying disease resistance within plant genomes. Tools like GPTransformers and dual-extraction modeling integrate genomic and phenomic data, boosting accuracy.

The study also notes challenges, especially the lack of high-quality, standardised datasets. To address this, it promotes federated learning—where models are trained across decentralised data sources without compromising privacy. Explainable AI tools such as SHAP are also emphasised to improve trust and transparency.

Ultimately, the study envisions a smart breeding pipeline combining multi-omics data, federated AI, and real-time field decision systems—marking a pivotal step toward securing the future of global food production.