Abstract
Crop diseases threaten yield losses, food security, and livelihoods, especially in Kenya, where agriculture is a signif-icant industry. Early detection and treatment of these illnesses can prevent crop losses, yet traditional disease de-tection methods are often inefficient, reducing agricultural yields and costing farmers money. This study explores the use of artificial intelligence (AI) to detect potato plant diseases through image classification to prevent losses.
The study evaluates the accuracy of an AI tool that uses Convolutional Neural Networks (CNN) for disease de-tection and classification. It also outlines the current state of potato disease detection and the integration of AI, considering challenges such as limited access to technology, insufficient training data, low technical knowledge among farmers, and potential disregard for indigenous agricultural knowledge. Recommendations to address these challenges include expanding training and capacity-building for farmers, facilitating access to technology and infrastructure in rural areas, availing quality training data, integrating indigenous knowledge into agricultural technology design, and establishing AI regulations.
The study also examines legal and policy considerations, including privacy, security, accountability, and AI regula-tory frameworks, emphasizing the importance of protecting personal data and ensuring transparent and account-able technology use. By addressing these challenges, potato farmers can effectively use AI applications for early disease detection, preventing plant losses and improving yields and food security in Kenya.
