Plans to deploy artificial intelligence in the agricultural sector are becoming increasingly concrete. But is that wise?

Imagine a wheat field stretching to the horizon. The wheat will be ground into flour, from which bread will then be baked to feed the people of the city. Everything in the field – planting, fertilizing, monitoring and harvesting – is controlled by artificial intelligence (AI). Think of algorithms that control the self-propelled tractors and combines, respond to the weather and meet the exact needs of the crop. It sounds like the future. However, this currently still has major drawbacks. Because what if, for example, a hacker messes up?

No science fiction

This illustrated example may still seem like a future. But according to researcher Asaf Tzachor, we should not be mistaken. “The idea that intelligent machines run farms is anything but science fiction,” he says. “Large companies are already pioneering the next generation of autonomous ‘ag-bots’ (agricultural robots, ed.) and systems that can replace farmers. But so far, no one seems to be wondering whether there are risks associated with rapid deployment of artificial intelligence.”

Use of artificial intelligence in agriculture
Several ‘ag-bots’ and advanced machines – such as drones and sensors – are already being used to collect information about crops and help farmers make choices. Artificial intelligence can help detect diseases and determine whether there is (inadequate) irrigation. Self-propelled combines can also bring in a crop without the need for a person. Such automated systems aim to make farming more efficient, save labor costs, optimize production and minimize loss and waste. This leads to higher incomes for farmers, but at the same time to a greater dependence on computers.

Despite the fact that artificial intelligence can make an important contribution to an improved agricultural sector and agricultural productivity, the potential dangers of this should not be forgotten, the researchers warn. In their study, they argue that the future use of algorithms in agriculture poses significant risks to farms, farmers and our food security. And these risks are currently still poorly understood and undervalued.

Cyber ​​attack

In the new study Researchers have listed the potential risks of implementing artificial intelligence in agriculture – and ways to address them. One of these possible dangers is an unforeseen cyber attack. Hackers can paralyze commercial farms or disable sprinklers, autonomous drones or ‘robot harvesters’. To guard against this, the researchers propose that ethical hackers help companies detect any security flaws during the development phase, so that systems can be protected from real, bad hackers.

Environment

Another risk that should not go unmentioned are possible environmental consequences. AI systems are programmed in such a way that they ensure the best crop yield. But if they’re fixated on just that, they may be able to ignore the negative environmental impacts. This could eventually lead to overuse of fertilizers and soil erosion. In addition, over-application of pesticides in pursuit of high yields can seriously damage ecosystems and pollute waterways. The authors therefore suggest involving ecologists in the design process to ensure that these scenarios are averted.

Desirable

Despite the possible risks, further implementation of artificial intelligence in agriculture on the other hand is also desirable. With an estimated two billion people suffering from food insecurity – including some 690 million undernourished mouths and 340 million children with nutrient deficiencies – artificial intelligence technologies and precision farming can provide a solution. Especially in light of the current climate change and growing world population, smart technologies and algorithms can solve existing problems.

The researchers therefore do not argue that we should not apply artificial intelligence in the agricultural sector. However, we should thoroughly test new technologies in experimental environments to ensure that they are safe and secure. “Artificial intelligence is being hailed as the way to revolutionize agriculture,” said researcher Seán Ó hÉigeartaigh. “As we deploy this technology at scale, we need to closely monitor potential risks and aim to address them early in the design phase.”