Artificial Intelligence for Agriculture and Food Systems Innovation Research Network
Innovation
Kenya
Ethiopia
Uganda
About
Africa’s population is projected to reach 2.6 billion by 2050, necessitating a 70% increase in food production to meet demand. This poses significant challenges for agriculture and food systems amid resource scarcity, climate change, and socioeconomic pressures. Artificial intelligence (AI) offers transformative potential to address these challenges by driving innovation and sustainable solutions.
This initiative establishes and supports an innovation research network on AI for agriculture and food systems in Africa. In the first phase, the network comprised of 10 African-led research projects focused on developing, deploying, testing, and scaling responsible AI innovations to tackle key challenges in agriculture. The initiative also seeks to inform African and global AI policy and practice by leveraging insights from these projects.
In its extended phase, the initiative will scale its efforts to deepen its impact on agricultural innovation and sustainable food systems across the continent.
Subprojects
In a bid to increase food production amid growing drought concerns, this project intends to study the feasibility of using emerging technologies like IoT and AI on existing farming practices and develop sustainable strategies increase efficiency and effectiveness of agricultural production.
Common beans and Irish potatoes are among the most important food and cash crops for small scale farmers in Tanzania, but their yield is threatened by four diseases. Cutting off diseased leaves and plants help curb the spread of these diseases, making early detection important. This project aims to develop an ML model that will be able to detect the diseases earlier based on leaf imagery data and enable the farmer to make the appropriate decision for managing the spread of the diseases.
In spite of agriculture being the largest and most important sector of the Tanzanian economy, many farmers and other sector stakeholders face considerable challenges in increasing their yields. This is because they struggle to access economically viable technology. This project will use ML techniques to learn useful patterns from input data to provide models that farmers can use to predict crop yield or offer recommendations based on the seasons.
Nsukka yellow pepper is one of the varieties of pepper grown in Nigeria and its popularity has attracted stakeholders to improve and sustain its production. However, this production is faced with serious challenges like pest attacks, high costs of inputs and low profit margins due to middlemen in marketing. To help overcome these bottlenecks, this project aims to leverage AI tools and applications to collect datasets for early pest detection, provide support for early soil nutrient loss detection, improve water conservation and improve access to the market, thus increasing its value chain.
This research proposes to tackle one of the most challenging problems in agriculture; the detection and diagnosis of crop disease in the field by using low-cost smartphones with embedded assisted technologies. The goal is to elevate early crop disease and pest surveillance and diagnostic capabilities in the hands of cassava smallholder farmers at scale.
The Kenya Horticulture sub-sector is the largest in agriculture contributing 33% of the agricultural GDP. The threats of climate change have, however, affected both the productivity and profitability of the sector. Increasing temperatures and changes in atmospheric moisture have resulted in the emergence of new pests as well as an upsurge of existing ones. Despite its importance, tomato, an important crop in Kenya, is constrained by pests and diseases accounting for 80-100% losses, with the most common pests being tomato leafminer (Tuta absoluta) and white flies. This research intends to develop an AI-based spatial tool for the monitoring and surveillance of Tuta absoluta and whiteflies on tomato crops in Machakos County, Kenya.
Smallholder farmers who produce more than 70% of the food consumed in Kenya lack access to localised productivity maximising data and are extremely incapacitated by severe changes in weather. AI and IoT tools can offer a solution to these farmers, coupled by the rapid development of low cost devices that can support these technologies. Using an IoT-based tool with enhanced AI features and plugins, this team intends to test the viability of strengthening local economies through the provision of access to vital weather data that would enhance crop yield.
Irrigation has been one of the leading solutions to climate change and population growth challenges to food security in sub–Saharan Africa. However, it has failed to live up to its potential. There is a need to build institutional arrangements that integrate the public and private sector players in AI soil moisture and nutrient monitoring tools in Malawi. This project’s goal is to scale up sensor technology for soil water and nutrient monitoring in irrigated agriculture. It is expected that a public-private partnership will be set – up to be developing and deploying AI tools for monitoring soil moisture and nutrient in irrigated farming covering 660 farmers at 30 irrigation schemes.
In sub-Saharan Africa, losses incurred as a result of non-adapted agricultural practices to climate change are 30% (plant health management, fertilisation, irrigation) and the associated post-harvest losses are estimated at 4 billion US dollars. With the development of Climate Smart Agriculture, the main objective of the Tolbi AI project is to use a combination of Artificial Intelligence, satellite images, and local languages to provide small-scale agriculture producers and inform national agricultural policies with real-time information on yield forecasts using a field management platform that monitors plant health, fertilisation and water needs.
The community around the University of Energy and Natural Resources in Ghana is highly dependent on agriculture for their livelihood. However, due to scarcity of land, the farmers cultivate crops on small pieces of land and these crops are regularly infested by pests and diseases. This project aims to develop a deep learning based mobile and web app to efficiently detect cassava, maize, tomatoes and cashew pests/diseases. Due to high illiteracy rates in the farming communities, their AI system will be user-friendly and have a text-to-voice facility to communicate the results and recommendations in English and the popular local language “Twi”. This is to also facilitate easy usage by the visually impaired.
Organisation(s)
African Technology Policy Studies Network
International Centre of Insect Physiology and Ecology
KUMASI Hive
Principal Investigator(s)
Nicholas Ozor