Alexa, How Should We Farm?

At a tech expo this year, Microsoft claimed to know me better than I knew myself. Artificial Intelligence (AI) could pick my ideal ice cream flavour. But why do we need artificial intelligence to make our food choices?

Waiting in line, I hoped I would get strawberry. I got strawberry.

Admittedly, there were only three available flavours, and I didn’t have particularly strong feelings against the other options, chocolate or lemon. But why let AI choose at all?

What is artificial intelligence anyway?

Artificial intelligence is the science of training machines to perform human tasks. This can be done through ‘machine learning, ' where a system is programmed to look at examples of data, recognize patterns, and come up with conclusions. Systems get feedback, refine, and then repeat this process. For Microsoft’s ice cream selector, a camera read my face as I watched a selection of images – a beach, a tree. My face showed my emotions, my age, and my gender. The end choice – strawberry!

I might know what ice cream I want. But the world is huge and full of variables. An AI system can swiftly assess vast amounts of data and develop solutions.

Artificial intelligence improves farming

AI is being applied from farm to table. There are smartphone apps like Tumaini, an AI-powered app to help protect the world’s most popular fruit. Banana farmers need to quickly recognise signs of disease to stop the spread. Tumaini is a smartphone app trained on over 20,000 images of major diseases and pests that allows banana farmers to manage these issues easily and affordably. The developers say the app could be particularly useful in low-income countries and help to track crop disease across the globe.¹

Some farmers already use sensors, satellites, robots and drones to monitor their crops. AI can process data from these machines and adjust farming practices to suit, like adding less or more fertiliser. This is part of the new era of ‘precision farming’. Similarly, vertical farm company Square Roots is using AI to measure and monitor growing crops by analysing pictures of growing plants and adjusting environmental conditions to suit. "The operating system is our central nervous system. There are millions of data points," said founder Irving Fain.² "Artificial intelligence is constantly learning and predicting how to produce the best quality product."

Microsoft tried to read my feelings from my face. Elsewhere, UK scientists are developing an AI to read pig faces to assess their welfare – pigs can communicate certain emotions through facial expressions.³ Similarly, a Dutch company is using AI to monitor dairy cow behaviour to boost productivity by a reported 30 percent.⁴

Growing better flavour with artificial intelligence

Can AI tech even find original solutions to farming-flavour problems? Crops grown for quantity tend to show a decline in flavour, a phenomenon known as the ‘dilution effect’. US researchers are investigating ‘flavour-cyber-agriculture' as a way around this.⁵ Using AI to analyse growing basil plants, they discovered that 24-hour light generated the highest concentration of aromatic compounds. They had made the world’s first super-flavour basil. “Unless you’re in Antarctica, there isn’t a 24-hour photoperiod to test in the real world,” said co-author John de la Parra. He added that AI could use vast datasets of detailed agricultural information to improve food crops, faster than had ever been possible before.⁵

Driving down food waste with artificial intelligence

In factories today, much of the work is done by machines. But machines are programmed to make the same product, while fruit and vegetables naturally vary in shape, so imperfect produce is often rejected. The UK's Institute of Mechanical Engineers estimates that up to half of all the food produced in the world – equivalent to two billion tonnes – ends up as waste each year.⁶

But by making speedy, intelligent decisions, AI could reduce waste, such as identifying if a potato is better shaped for wedges or French fries. “We are talking millions and millions of tonnes of product being saved, optimum use of food and maximum yield from farm to fork," said Pieter Willems, technical director at TOMRA Systems, a Norwegian company focused on recycling solutions.⁷

Human brains vs artificial intelligence

Every second of every day, our brains calculate and combine sensory data and come up with conclusions. It’s still humans who manage agriculture and food production, who design food products, and who make grocery choices at the supermarket. But agriculture needs to become more efficient to feed the future world population. Companies need to cut down on food waste and energy costs. And there’s the rise of the Internet of Things – the interconnected system of devices, sensors and machines that transmit and share data – in the everyday world. Could it now be time to let AI do some thinking for us?