As AI gains traction in the agriculture sector, ethics questions need to be front and center.
Image Credit: istock
Artificial intelligence, or AI, has made inroads into almost every industry and socioeconomic sector in recent decades, with a promise of making life easier and more efficient. Smart home appliances keep refrigerators stocked and temperature controlled, while social media apps connect people to networks and online shoppers find just what they need with targeted ads. But these AI applications have also introduced new concerns for privacy, implicit bias and a host of potentially harmful, unintended consequences.
To head off some of these consequences, researchers have increasingly focused on studying the potential implications of AI across a variety of sectors. New analysis from the University of Maryland tracked the rising trend in such studies, and revealed that some sectors, like agriculture, have barely scratched the surface to identify and address potential ethical, legal, social, and economic implications of AI.
According to Debasmita Patra, an assistant research professor in the Department of Environmental Science and Technology at UMD and lead author of the study, agricultural businesses are relative late-comers to AI technology. But AI-dependent technologies are emerging rapidly in the agricultural sector, and Patra says now is the time to address these potential concerns.
“When a technology is new, we don’t know all the potential impacts of that technology,” Patra said. “And we need to be asking these questions about the ethics involved in AI applications in agriculture now, while it is still new and rapidly emerging, before we get very far down the road.
Patra and her colleagues conducted a global analysis of all the research papers focused on ethical, legal, social, and economic implications of AI from 1991 through 2020. Their study found that the greatest number of research papers on these issues were focused on the fields of computer sciences and social sciences—both fields in which the use of AI was adopted early and has grown rapidly. They also found that fields like healthcare, where ethical issues are often front and center, were the focus of a large number of studies on the implications of AI.
But the team saw a drop off in the number of such studies where AI has been slower to make inroads, such as agriculture. Of 1,028 total studies analyzed, only 23 explored the implications of AI use in agriculture.
The agricultural sector is rapidly developing more uses for AI, and each new application poses many unanswered questions. For example, satellite imagery of farm fields and farmer-collected data on leaf moisture can be used by AI systems to predict disease or pest infestation and recommend management practices, but once collected, who owns the data?
Farms are not islands, and conditions on a large farm may provide important information for managing a broader region. Does that make it OK for agribusinesses selling AI-based management services to farmers to use data from those farms in other land management applications? Or can technology companies use farm data they collect to try to upsell products to the farmer, or recommend a particular brand of seed or supplies that offer them a kickback?
Other concerns include how sustainability will be factored into AI applications designed to increase yield. For example, an AI-driven application may recommend measures to improve harvest that also cause more nutrient loss to surface water run-off and pollute the surrounding habitats.
There is also the potential for AI-guided automation to lead businesses to modify their livestock or crops to better suit a robotic tender or harvesting machine. Equity is also an important consideration, as AI technology is expensive, and not all farmers will reap the benefits. These are just a few of the issues that AI is introducing into agriculture, and Patra and her colleagues expect research in this area to continue as AI adoption grows.
The research paper, Ethical, legal, social, and economic (ELSE) implications of artificial intelligence at a global level: a scientometrics approach, was published January 13, 2022, in the journal AI and Ethics.
Other co-authors of this paper from UMD include: professors Adel Shirmohammadi from the Department of Environmental Science and Technology (ENST) and Abani K. Pradhan from the Department of Nutrition and Food Science (NFSC) and the Center for Food Safety and Security Systems; undergraduate students Madeline White and Christopher Broussard from ENST, Joel Cover and Liana Torres from the Department of Environmental Science and Policy; graduate students Edmund Benefo from NFSC; and Aubrey Tingler from the Marine-Estuarine-Environmental Sciences Program.
This study was the result of research conducted by undergraduate students, graduate students, and faculty from multiple disciplines during the undergraduate and graduate level courses, Ecological and Natural Resources Ethics ENST404 and Advanced Natural Resource and Environmental Ethics 604 taught by Debasmita Patra (pictured on the left).
“Besides its timely scientific contribution to the body of literature, this study was innovative in bringing classroom teaching efforts into publication stage,” said Adel Shirmohammadi, who is a professor in ENST and a co-author of the study.
“The combined effort of undergraduate and graduate students, and faculty from diverse disciplines indicates an innovative teaching approach that traversed from literature review to data synthesis and analysis and finally to publication in an appropriate Journal,” said Abani Pradhan, a professor in NFSC and co-author of the study.
This work forced me to pause and rethink, is there a better/ethical way of integrating AI in agriculture as I attempt to use AI in my Ph.D. research?Ednumd Benefo food science graduate student
Working on this research in the ethics class was such an exciting and unique experience for me. It was wonderful to work with such a hardworking group of people and I’m so glad to see all of our work finally pay off with a publication.Madeline White undergraduate student