Written by Derin Sezgin, March 30th, 2025
In 2022, 735 million people faced chronic hunger, not consuming enough food to meet the body’s requirements for healthy living. In total, a staggering 2.4 billion people faced “moderate to severe food insecurity” (lacking sufficient nutrition on occasion to commonly), which is a number escalating rapidly compared to 2019, with nearly 400 million more people facing malnutrition (United Nations, Global Goals: Goal #2). Clearly, the previously substantial problem of hunger does not seem to get better at all and, instead, is in immediate need of a solution that will turn the tide for about 30% of the people of this world. The goal of accomplishing zero hunger by 2030 (set by the United Nations in 2015) is now just a distant fantasy, and the world must set more pragmatic goals.
In search of transformative solutions, scientists, aid organizations, and policymakers have turned to Artificial Intelligence to solve this issue. But can AI truly bring agriculture to its peak effectiveness, or will it cause minimal change and significant financial burdens, deepening this problem?
To begin with, artificial intelligence has a lot of potential to aid significantly in preventing hunger with its advanced data management capabilities. The world already has a strong foundation: remote sensing, earth observation, and satellite imaging. These are just a few analytical tools already in place that could help optimize agriculture. But now, with the advent of Artificial Intelligence, this can all be automated and made more efficient, bringing data to farms worldwide. If you’re wondering how AI will play a role, well, here it is: “Land use decisions will benefit from machine learning to more accurately determine the most suitable plots and crops based on climate modeling, ecosystem data, and disaster risk mapping” (Artificial Intelligence Can Transform Global Food Security and Climate Action, Tshilidzi Marwala). This technology would significantly reduce the risk of losing a significant crop harvest due to something unexpected, as now, it is possible for the world to expect a natural disaster, drought, or anything similar. Not only that, machine learning and artificial intelligence will also be capable of informing agricultural estates of prominent details like soil health and water availability, which are crucial to understanding why some plots might underperform.
These methods are up-and-coming; they have the genuine potential to reverse the landscape of global hunger completely, but they are only as good as the quality of data that fuels them. Consider this: AI has successfully optimized agriculture across the world (which means it was made the most efficient possible, as farming has been ‘optimizing’ since the Fertile Crescent 12,000 years ago), and farmers have enough food and crops to go around to everyone who needs it to lift themselves out of a situation of constant malnutrition. Now, how do we ensure that the crop goes around to everyone who needs it evenly and in time, as fresh produce is always at risk of expiring? This is where new potential uses of AI come in. Using ML (Machine Learning), we could track the rough expiry dates of crops based on how the typical produce looks at every point during its life cycle (from freshly picked to expired). Bernhard Kowatsch, the head of the World Food Programme (WFP)’s global innovation accelerator, thinks the world must “consider precision agriculture or personalized nutrition for everybody, or AI optimized collaborative systems that ensure farm-to-fork digitization on blockchain to enable more food, higher nutrition, and lower costs all at the same time (Bernhard Kowatsch, Can AI solve hunger? The promise of technology in food security). Solving hunger worldwide seems to have two facets: optimizing crop (food/produce) production and agriculture and distributing the food evenly, efficiently, and timely. With the AI solutions discussed in the previous two paragraphs, the hunger problem looks like it can and will be solved soon (in the next 20 to 30 years).
Unfortunately, integrating artificial intelligence into hunger prevention is not easy; these AI solutions have flaws. Firstly, there could be a matter of security concerns regarding Artificial Intelligence-based agricultural tools. There could be scenarios in which these AI systems, susceptible to malicious attacks, face significant problems and significantly reduce their crop production efficiency and rate, leading to negative progress in alleviating hunger in the area. These attacks could also become frequent, and one scenario strikes my mind: agricultural estates competing in the same area to become the prevailing producer decide to hinder their opponents’ progress by cyberattacking them. For example, “the 2021 cyberattack on JBS, the world’s largest meat processor, foreshadows [these] potential risks.” The company was “forced to pay around $11 million in ransom to end the attack on its computer networks, which temporarily shut down some of its operations in Australia, Canada, and the US” (Oliver Morrison, Risks of using AI to grow our food ‘poorly understood’ and ‘under-appreciated’). To establish the necessary protections to ensure this would never happen, major producers would have to hire ‘white-hat hackers,’ or cyber-security experts that aim to protect the AI systems from malicious software. This would also add extra costs for companies that make this AI-integrated agriculture less and less affordable.
Furthermore, the use of autonomous machines and artificial intelligence could further the gap between subsistence (farming to maintain a living) and commercial (farming to make a profit) agricultural estates, as well as a significant problem in the world’s agricultural workforce. First, the estimated cost of an AI-controlled farm, as described previously, at a small scale is well over $75,000 (considering the AI-powered robots, maintenance costs, and energy costs). Currently, “84% of the world’s 570 million farms are smallholdings; that is, farms less than two hectares in size” (Hannah Ritchie, Smallholders produce one-third of the world’s food, less than half of what many headlines claim). Given that two hectares equal 0.02 square kilometers, a cost of $75,000 is nearly impossible for small-scale subsistence farmers who can barely afford a living. Additionally, “marginalization, poor internet penetration rates, and the digital divide might prevent small holders from using advanced technologies” (The University of Cambridge, Risks of Using AI to Grow Our Food Are Substantial and Must Not Be). Suppose larger-scale, commercial agricultural estates gain access to this technology and increase efficiency and crop production. In that case, it will push nearly 9/10 (84%) of the 2 billion people worldwide into this industry (International Labour Organization, Employment in Agriculture (% of total employment)) into a position far below the poverty line in a position where they might not be able to afford living necessities like food and shelter. Not only would this lead to a significant spike in unemployment worldwide, but it would also create a complete crisis in hunger alleviation. If most of the farmers who produced the food that the world relied on transitioned from producers into desperate consumers, then the entire world would have a challenging time keeping their stomachs full.
This journey has taught us that integrating AI systems into agriculture is not a seamless solution at all, as its implications can lead to outstanding crises. But, like any other problem in the world, those implications of cybersecurity and an unemployment and hunger crisis can be avoided through the protection of agricultural AI systems and the smart and slow integration of AI into agriculture, with protection and workforce re-integration plans for farmers that might lose their jobs due to these systems. We stand at a critical juncture; will we take action to embrace and integrate sustainable AI agriculture for a better future, or will we turn our backs on the growing crisis of hunger?”
Leave a Reply