Smarter farming is a moral imperative, and big data provides the tools

Opinion

 
by

A few decades ago, the world figured out that food production was not meeting the demands of population growth, and the Green Revolution was born. The idea behind it was that by providing farmers with a ‘package of practices’, including good advice and good seeds and other inputs, on-farm productivity would rise. This fundamental strategy is still true, but the reality is that the vast majority of farmers worldwide still do not have access to these technologies.

The CGIAR’s Standing Panel on Impact Assessment has begun to look at the real rates of adoption of on-farm advice, and has found a gap between rhetoric and reality. Overall, the area of arable land is declining, at the same time as populations are growing and climatic conditions are becoming more variable; in order to meet the food security challenges of future generations, we need to do things differently.

Agriculture is nested in a complex web of interactions between farming, ecology, society, economy, and technology. Production decisions must be informed by all of these areas of knowledge; this is where big data and ICTs open up new ways of working. Digital companies are very good at tailoring experiences to individual consumers through methods like user-centred design (i.e. advice for your farm for the season over a whole landscape or agricultural zone), and in running many thousands or millions of user tests to make digital media more engaging and interactive (A/B testing). My view is that these approaches are the natural next step for linking agronomy to the impacts and adaptability we need to build resilient global food systems.

Agricultural services at scale

In terms of digital innovations, I get a lot of inspiration from start-ups like Farm.ink and WeFarm, who are beginning to show us the power of farmer-to-farmer digital communities, and to provide ways for subject matter experts to interact directly with farmers at scale. Over time, I think we will see increases in both skill (as reflected in crop outcomes impacting household income) and adaptability (ability to quickly change farming practices by year and location) in ways that the world sorely needs.

One ongoing project we have with Farm.ink links a Facebook Messenger community of over 150,000 farmers (and growing) with bilingual experts from the International Livestock Research Institute to build up farmers’ animal health and dairy-producing capabilities. A survey found that over 90% of the farmers had adopted one or more improved practice. Being able to reach the community online serves as a dynamic way to provide up-to-date information in animal health and disease to farmers across a large scale.

Digital dangers

Although there are examples of impactful projects, agriculture is probably the least digitised economic sector, whereas finance is probably the most digital. If we want to glimpse future dangers of digital to the ag sector, we can start by looking at finance. The automation of large swathes of the financial sector is creating an environment where ‘predatory algorithms’ could inform the development of ‘predatory loans’ targeting the poor. We could imagine such a ‘dark side’ of the kinds of analytical work we do as researchers. For example, we commonly generate agro-climatic forecasts for regions, crops, and even individual farms that would be of great value to the financial sector for developing new pro-poor advisory services and financial products. If we are not careful, these analytical products could also inform predatory lending.

At the CGIAR Big Data Platform, our first job is supporting researchers to understand what is at stake with regards to responsible management of farm and farmer data, and to observe responsible data practices throughout the research data lifecycle. Last year, we reviewed about 150 frameworks and other resource documents to distil information regarding responsible practices into some very actionable guidelines. But we need more than guidelines; we need what we might term ‘technologies of privacy’ that will enable us to protect farmer data even while conducting analysis and communicating with farmers.

That being said, we recognise that the environment is changing very quickly, and ethical frameworks struggle to stay abreast of the rate of change. Ethically, we are required to inform farmers of all the ways we would anticipate using their data and secure their consent and yet, sometimes, the notion of truly informed consent is complicated by new methods and learnings. At our upcoming Big Data in Agriculture Convention, we’ll propose a more dynamic approach to consent built on continuous interactivity and partnership with farmers that we hope will help us navigate this complexity while securing the benefits of digital services, and personalisation at scale.