Advances in Artificial Intelligence in Crop Breeding: Opportunities, Challenges, and Future Directions
Year: 2026
Type: article
Abstract: Artificial Intelligence (AI) is increasingly recognized as a transformative tool in crop breeding, offering predictive, data‑driven solutions to accelerate genetic gains. Traditional breeding approaches remain constrained by long selection cycles, limited capacity to capture complex genotype × environment interactions, and the urgent need to address climate change and food security. The objective of this review is to synthesize recent advances in AI applications, including machine learning (ML), deep learning (DL), and generative AI, in genomic selection, high‑throughput phenotyping, and decision support systems, while identifying research gaps and outlining future opportunities. Recent results demonstrate the tangible impact of AI in breeding pipelines. Deep learning models have improved genomic prediction accuracy for yield in maize and wheat by up to 15% compared with conventional statistical methods. Computer vision approaches applied to unmanned aerial vehicles (UAVs) and hyperspectral imagery have achieved disease detection accuracies exceeding 90% in rice and soybean, while reinforcement learning frameworks are beginning to optimize parental selection strategies under variable environments. Generative AI has also been applied to augment small phenotypic datasets, enhancing model robustness and enabling more reliable predictions. Despite these advances, challenges persist. Current models often rely on heterogeneous and incomplete datasets, limiting reproducibility across crops and environments. The “black‑box” nature of deep learning restricts biological interpretability, and computational demands hinder adoption in resource‑limited breeding programs. Future directions include AI‑guided genome editing for targeted trait improvement, federated learning for cross‑institutional collaboration without data sharing, and digital twins of crops to simulate growth under diverse scenarios. Addressing these gaps will enable AI to deliver climate‑resilient, nutritionally enhanced, and sustainable crop varieties for global food security.
Source: Journal of chemical health risks
Institution: Gandhigram Rural Institute
Language: English
Topic: Smart Agriculture and AI
Subfield: Plant Science
Domain: Life Sciences
Keywords: genomic selection, high-throughput phenotyping, machine learning, deep learning, generative AI, computer vision, unmanned aerial vehicles, hyperspectral imaging, plant disease detection, reinforcement learning, parent selection, gene-environment interaction, data augmentation, neural network interpretability, federated learning, digital twin, climate-resilient crops, food security
SDG: Zero hunger
Open Access status: diamond
OpenAlex ID: W7217941367
