Big data is coming to agriculture. Farmers must set its course

The industrialization of agriculture began some 100 years ago. We are now witnessing its digitalization. But a wave of big data may sweep farmers off their land, unless they mark out a course in good time and decide which problems digital technologies should address.

In the future, young farmers are likely to don digital glasses or consult other devices that will help them analyze their work and make decisions. Their data might be collected by self-piloted multicopters, which review the state of the field according to empirical formulas and provide specific, effective cultivation tips for individual plants, for both organic and conventional farming.

A bit more manure here, some of the latest insecticide there? Should the tomato be deprived of water a tiny bit longer so that it develops the perfect flavour? Does Daisy the cow's temperature indicate that insemination should wait until the afternoon? And can the purchase of feed pellets be put off until next week, after the markets have calmed down?

These are the questions that agricultural "Siris" are already beginning to answer, even if they are still in the pilot phase. For example, a six-legged robot named Prospero is roaming test fields in the US, planting individual kernels of corn in exactly the right spot for the plant to take root. And for several years now, Bonirob has been wandering the fields of Germany unassisted, testing the ground and picking weeds that threaten the main crop.

You don't have to be a clairvoyant to recognize that agriculture is undergoing rampant digitalization. Automation is as inevitable as the tasting of the forbidden fruit. The promises of technology are too seductive, and the lure of greater efficiency all too tempting.

Shouldn't we proceed with caution? Just because we want something, it doesn't mean that it's good for us. Diversification and variety trump everything, especially when it comes to agriculture and food. Simplistic, cookie-cutter approaches to solving problems usually reveal considerable weaknesses, fast. Food production is a highly complex endeavour: millions of organisms in a single litre of soil affect the development of the crops that grow in it. Likewise, thousands of compounds in the plant affect the cow that eats it. We cannot package everything into a single correct formula. But is that a reason not to devise any formulas at all?

I don't think so. Rather, it's more important to ask: what new formulae should we create? Should digitalization aim to reduce costs in the short term, or preserve environmental resources in the long term? Which "apples" should it make tastiest? Compared to the rest of the world, our agricultural practices are very sustainable, whether they are organic or not. And our farmers are highly knowledgeable and competent. This is why I think we should get involved now, and decide which problems future technologies should tackle.

What will be left for farmers to do when agriculture is digitalized and automated? Will they be reduced to servants of algorithms and of machines that require a few remaining manual manoeuvres, or will they take on a new role? I think farmers will primarily assume the role of technically skilled researchers. New diseases will emerge, organisms will migrate, and unusual cases and technical problems will arise. Yes, farmers may no longer steer tractors themselves, but they will still have to check the fields and stables, take care of specific tasks and improve systems interactively on an ongoing basis.

In the field of medicine, technical advances have still not supplanted doctors or nurses. Instead, they have made it possible for medical staff to deal with more complex diseases, as we begin to live longer and often require long-term care. In agriculture, digitalization may in fact make it possible to focus more on the true heroes of the story: plants and animals. Or we will see a growing trend towards complexity and diversity. And there will be more time to tackle new challenges. Or more time to advise other farmers in remote places, who do not have the same access to education and technology as us.

Granted, the world doesn't actually work this way. But wouldn't it be fantastic if it did? We now have the opportunity to develop a model for what digitalization should bring to agriculture, and to set its course. This is preferable to standing back and waiting to see the toll it will take - both on farming and on us.

This article was first published in ETH Zürich's Zukunfstblog