By: Roberto Ramos-Diaz and Dr. Aaron F. Brantly
The global demand for food is growing. The World Bank estimates the caloric demand will increase alongside the population. The limitations facing the production of food to sustain a growing global population is forcing the utilization of new technologies in agriculture. Every day new technologies, in particular Artificial Intelligence (AI) and Machine Learning (ML), are being developed and implemented on farms in the United States and around the world. Yet, rarely is the changing nature of technology on farms considered from a human perspective. Specifically, is AI in agriculture, in fact, benefiting or harming humanity?
There is little doubt that we are in the midst of a technologically enabled agricultural revolution. Historically, societies progressed from surpluses to shortages as the state became increasingly developed. Analysis in The Future of Work in Agriculture as well as in the World Economic Forum characterize the present - Fourth Industrial Revolution - as being unprecedented. The speed at which this revolution is taking place is unpredictable yet unrelenting. Technologies in agriculture are constantly changing and improving through the use of AI and ML in tandem with ever-increasing quantities of data. The three previous revolutions consisted of technological changes that allowed for the automation of often simple and repetitive tasks. Earlier revolutions in agriculture were in response to farming labor shortages, rising cost of production, including wages, and or a need to raise labor productivity. Presently, the US agricultural labor force is heavily dependent on foreign-born laborers. These workers offset a decline in domestic workers willing to engage in agricultural work. In addition to declining domestic agricultural labor, populations in rural areas devoted to agriculture have been shrinking and aging.
Forecasting the future of agriculture in a time of change is difficult. Will a supply-side miracle of long-term gains in efficiency and productivity occur and can the labor force can adapt itself to the new skills required to manage, operate, and maintain advanced AI imbued machinery? A future of AI enabled combines and other equipment seems more suited young technologically savvy laborers. Or, will the agricultural AI revolution yield increasing inequality? In particular, will AI disrupt labor markets and create gaps between returns to capital and returns to labor such that the role of human labor becomes so small as to be inconsequential? Finally, is there a third option in which there is a balance between the elimination of labor and the transition to newer more advanced labor markets? This third – hybridized model balances the timing of implementation of technologies with the ability of labor markets to adapt.
AI, responsibly and effectively implemented, will likely result in increased harvest yields and reduced waste. Emerging market losses in agriculture are estimated to be approximately ⅓ or 1.3 billion tons. These losses occur during the production, post-harvesting handling, storage, and processing stages. The practice of farming as a whole can be made more sustainable and efficient. For example: the use of artificial intelligence aids in the reduction of fertilizer and pesticides, as a result of improved pest and disease detection. Moreover, AI can also facilitate automated crop grading (assessing crop quality). Precision resulting from AI will change agriculture market structures. AI utilization can also reduce the cost of agricultural insurance premiums through greater precision in developing risk assessments resulting in reductions in crop losses. Increased data flows improve access to new markets through the reduction of trade friction. Each of these advances help large and small farmers better integrate into regional and global supply chains.
Yet concurrent to changes in agricultural market efficiencies are externalities that impact the laborers that depend on non-AI enabled farming practices. For instance, remittances from migratory agricultural labor is a significant source of financial capital for those countries from which that labor originates. The loss of this capital due to the alleviation of labor market demand has cascading global effects well beyond farming communities where products are grown or raised.
As agricultural labor markets decline due to increasing automation, the communities that support and sustain that labor force also suffer. The result is a shrinking population, with diminishing services and fewer opportunities. As labor opportunities dry up rural food deserts in the midst of plentiful farmland become increasingly common. Here the impact of automation is not limited only to those activities that occur on the farm but instead create negative externalities that extend beyond and have both local and regional impacts.
Automation is not without costs to the farm owner either. Initial capital outlays necessary for procuring AI enabled equipment are steep and often require farmers to undertake increasingly substantial debt. There are few alternatives to the procurement of advanced agricultural equipment if farms wish to remain competitive. Yet, increasing debt also increases the potential for loss as farmers compete in global markets, subject to environmental changes, global economic shifts and political concerns well beyond their control. It is a catch-22 in that farmers who retain human labor are less efficient cannot compete, while those who automate must undertake increasing levels of risk.
What should be done about AI in Ag?
The choice to implement AI in agriculture is seemingly Fait Accompli. A paper published in the Proceedings of the National Academy of Sciences of the United States of America (PNAS) recommends viable worker retraining programs. Yet it is unclear whether these training programs can truly sustain a viable agricultural labor force. There are many roadblocks to retraining labor that includes both the willingness of workers to continually retrain into new and uncertain careers, and the provision of sufficient financial resources to even conduct training.
Fostering local food movements provides another area for potential labor equity and justice. By shifting from global to local markets some farmers may be able to increase labor participation and develop communities. While the local farm-to-table movements might provide a small offset to larger scale agricultural trends, they are unlikely to displace broad market shifts towards automation. Yet, these smaller community-based initiatives might offer a means to support local community dynamics in many areas.
As a nation and a world, we are already well down the path towards automation in agriculture. AI in agriculture is projected to grow from an estimated USD 1.0 billion in 2020 to USD 4.0 billion by 2026, at a CAGR of 25.5% between 2020 and 2026. Understanding the issues raised above and addressing them in a timely and equitable manner during this transition period will shape the future of the global economy and the livelihoods of those involved in the agricultural labor market. A failure to understand and address the impact of AI on agriculture and the humans employed in the growing and raising of crops will likely challenge social, political and economic structures at local, national, and global scales.