The risks of using AI to grow our food are ‘misunderstood’ and ‘underestimated’
Artificial intelligence promises to allow us to produce more food using fewer resources. Cost has been the biggest barrier to widespread adoption so far. But the risk of cyber attack, environmental damage, worker exploitation and widening the gap between commercial and subsistence farmers could all be unintended consequences of harnessing the power of AI to grow our food. .
These risks are significant and should not be ignored, the researchers warn.
A new risk analysis, published in the journal Intelligence of natural machines, warns that the future use of AI in agriculture carries significant potential risks to farms, farmers and food safety that are poorly understood and underestimated.
“The idea of intelligent machines running farms is not science fiction. Big companies are already at the forefront of the next generation of autonomous ag-bots and decision support systems who will replace humans in the field,”said Dr Asaf Tzachor of the Center for the Study of Existential Risk (CSER) at the University of Cambridge, first author of the paper.
“But so far, no one seems to have asked the question: are there any risks associated with a rapid deployment of agricultural AI?”he added.
Despite the enormous promise of AI to improve crop management and agricultural productivity, potential risks need to be responsibly addressed and new technologies properly tested in experimental settings to ensure they are safe and secure. against accidental failures.
Cyber attack alarms
First, the research is sounding the alarm about cyberattacks that could disrupt commercial farms using AI, by poisoning datasets or shutting down sprayers, autonomous drones and robotic harvesters.
The 2021 cyberattack on JBS, the world’s largest meat processor, for example, portends potential risks associated with the introduction of digital technologies into agrifood supply chains, according to research. The meat giant was forced to pay around $11 million in ransom to stop the attack on its computer networks, which temporarily halted some of its operations in Australia, Canada and the United States.
Likewise, last year’s ransomware attack by Russian hackers on NEW Cooperative, which provides feed grain to 11 million farm animals in the United States, further accentuates this emerging cybercrime landscape.
To guard against this, the researchers suggest that “white hat hackers” — or “ethical” hackers who use their skills to help companies improve their IT security — help companies discover any security flaws during the development, so systems can be protected from real hackers.
Food and agriculture technology is increasingly recognized for its transformative potential, the opportunity it opens up to develop a more sustainable food system that can do more while protecting and restoring natural capital.
However, the authors suggest that an AI system programmed only to deliver the best crop yield in the short term could ignore the environmental consequences of this achievement, leading to overuse of fertilizers and soil erosion in the long term.
Excessive application of pesticides in pursuit of high yields could poison ecosystems and excessive application of nitrogen fertilizers could pollute surrounding soil and waterways, they argued. The authors suggest involving applied ecologists in the technology design process to ensure these scenarios are avoided.
For example, if monocultures – where a single genotype of a plant species is grown over large areas – are irrigated, fertilized and inspected by the same suites of algorithms, a model error or poorly calibrated sensors can lead to a excessive fertilization and the soil microbiome. degradation, at the risk of large-scale crop failures.
Exploitation of workers
In addition to reducing the overall carbon footprint of food production, technologies such as plant breeding, precision farming, robotics and automation, vertical farming and mobile apps also promise to reduce reliance on farmers with respect to hired labour.
Autonomous machines can improve working conditions for farmers, relieving them of manual labor, the researchers observed. But without “inclusive” technology design, the socio-economic inequalities that are currently entrenched in global agriculture will persist, Tzachor warned.
“AI expert farming systems that do not take into account the complexity of labor inputs will ignore and potentially support the exploitation of disadvantaged communities,”he said.
Various advanced ag-bots and machines, such as drones and sensors, are already being used to gather crop information and support farmers’ decision-making: detecting diseases or insufficient irrigation, for example. And autonomous combines can bring in a harvest without the need for a human operator. These automated systems aim to make farming more efficient, reduce labor costs, optimize production and minimize losses and waste. This leads to increased income for farmers as well as greater reliance on agricultural AI.
However, small-scale producers who cultivate the majority of farms globally and power large swaths of the so-called Global South are likely to be left out of AI-related benefits, Tzachor pointed out. Marginalization, low internet penetration rates and the digital divide could prevent smallholders from using advanced technologies, thus widening the gaps between commercial and subsistence farmers, he argued.
Meanwhile, partial, biased, or irrelevant agricultural data can hamper ag-tech efforts to improve food security.
With an estimated two billion people suffering from food insecurity, including some 690 million people suffering from malnutrition and 340 million children suffering from micronutrient deficiencies, artificial intelligence technologies and precision agriculture promise substantial benefits for food and nutrition security in the face of climate change and a growing world population.
But the people and practices at the heart of indigenous agricultural systems are often underrepresented in the data, despite their contribution to local food security and dietary diversification. For example, typical agricultural data sets have not sufficiently considered polyculture techniques, such as tree farming and silvograzing. These techniques produce a range of food, fodder and textile products while increasing soil fertility, controlling pests and maintaining agrobiodiversity.
This is important as noted by the researchers “a decades-long focus on staple crops such as wheat, rice and maize has overtaken research efforts on crops of critical importance to poorer producers and subsistence farmers, notably quinoa, cassava and sorghum”.
“AI is being hailed as the way to revolutionize agriculture,” added Dr. Seán Ó hÉigeartaigh, Executive Director of CSER and co-author of the new research. “As we deploy this technology at scale, we need to take a close look at potential risks and aim to mitigate them early in the design of the technology.”
Responsible artificial intelligence in agriculture requires a systemic understanding of risks and externalities
Intelligence of natural machines