Africa is the birthplace of humanity, but in the 21st century there is still the sense that it is a continent largely left behind by the march towards economic development and modernity. Agriculture continues to be the largest contributor to the economies of most sub-Saharan African countries, but the sector is dominated by subsistence rainfed production, characterised by low productivity and underdeveloped market opportunities. Some of the major contributing factors include poor infrastructure, limited access to and availability of inputs, and the difficulty that farmers face in getting up-to-date and relevant information to improve their farm output.
Of course, these challenges do not mean that agricultural development programmes are not delivering. On the contrary, important successes are being realised in all areas of agricultural development. But the speed of change can be frustratingly slow. Technological innovation is having remarkable impacts on a wide range of human endeavours, including agriculture. Are there opportunities for Africa to fast-track the improvement of its farming systems through the targeted application of these technologies? I would argue that the answer to this question is, ‘yes’. My reason for this answer is three-fold:
- Leapfrogging technology – African countries have already proven to be very adept at overcoming infrastructural weaknesses by leapfrogging steps in technological development. Prominent examples include the adoption of mobile phone technology by users who have never had a cabled landline connection, and the use of mobile banking services by rural communities that lack even the most basic physical banking facilities.
- Rapid penetration of mobile communications technology – according to reports from the GSM Association, penetration of mobile subscriptions in Africa will increase from 44-52% between 2018 and 2025. More importantly for data applications, mobile internet usage will rise from 21-40% and smartphone adoption from 34-68% over this period.
- The emergence of artificial intelligence (AI) – the development of AI applications is moving into an almost exponential phase. The co-occurrence of this new technology with the on-going mobile revolution in Africa will mean that AI will be at the forefront of novel phone apps aiming to strengthen African agriculture.
Harnessing the potential of AI
With these points in mind, my team from IITA has joined forces with a team from Penn State University, led by David Hughes, to begin ‘walking the walk’ with AI for agricultural development in Africa. Together, we are developing a smartphone app that will use machine learning-based AI to diagnose cassava disease and pest damage. More cassava is produced in Africa than any other crop, and it plays a critical role in safeguarding food security, but it has been plagued by virus diseases for decades. In fact, more than half of all cassava plants in Africa are affected by two virus diseases: cassava mosaic disease (CMD) and cassava brown streak disease (CBSD).
CMD and CBSD result in losses of over €800 million every year in Africa. The diseases both have symptoms that can be clearly recognised by trained researchers or extension workers, and, once identified, good control advice is available. However, very few farmers know about the diseases, and even fewer still have access to control information. This problem set myself and the other researchers at Penn State and IITA thinking, what if farmers could recognise these diseases themselves? What if they could also get advice on how to control them? And, what if they could be linked up with a network of growers, agricultural workers and researchers all working towards controlling the diseases and improving cassava productivity?
Scaling adoption of the app
Trials of the AI app in Tanzania are providing highly promising results: the app can distinguish between CMD and CBSD, as well as cassava green mite damage and healthy leaves. The app will also provide farmers with control information and links to nearby extension/research officers, as well as other sources of information. Initially the app will be released to a limited number of targeted users to obtain final validation data. However, it is hoped that a beta version will be rolled out before the end of 2018. This should be freely downloadable and usable offline by any person with a smartphone.
The team plans to promote the app and encourage farmers and extension workers to use it through an extensive training and awareness-raising effort planned for 2019. A key aspect of this will be demonstrating how the app can be used to help farmers access disease control measures – such as improved seed varieties and disease-free planting material – as well as to apply basic phytosanitary measures, including the selection of healthy stems for planting.
The AI app is one small example of the information and farming revolution that is coming to Africa very soon. The Penn State/IITA team is hugely confident about the potential for the cassava AI app, but this will surely be just one of an increasing number of phone-based software tools that will help to raise hundreds of millions of African farmers out of poverty.