UESC researchers launch AI app to assist cocoa farmers

Researchers from the State University of Santa Cruz (UESC) in Ilhéus, Brazil, have developed an artificial intelligence-powered application designed to assist cocoa producers in identifying agricultural issues. Launched during the second Expocacau event, the tool acts as a virtual assistant that can process text, audio, photos, and videos to provide preliminary diagnoses of pests, diseases, and nutritional deficiencies.

A primary focus of the tool is the detection of ‘vassoura-de-bruxa’ (witches' broom), a major fungal disease affecting cocoa crops. The project aims to support small-scale farmers who often face delays in receiving technical assistance, which can lead to significant financial losses if diseases are not identified quickly.

The development is an international collaboration involving researchers from the Universidad Nacional de San Martín in Peru, who provided training images, and Arun Pratihast from Wageningen University in the Netherlands, who developed the AI model. The project is coordinated by Dr. Deborah Faria and is funded by the Instituto Arapyaú through the South Bahia Science and Technology Park (PCTSul). Future updates are expected to include features such as crop yield estimation.