Socially Responsible AI Assurance in Precision Agriculture for Farmers and Policymakers

Posadas, Brianna B.; Ogunyiola, Ayorinde; Niewolny, Kim

As one solution to feeding a growing population with finite resources, some farmers, researchers, and agricultural technology providers (ATPs) have turned to precision agriculture (PA). PA is the practice of mapping out precise input application to maximize the yield. To do this, ATPs collect input and output data from farmers and use Artificial Intelligence and machine learning to build prescription maps, which farmers can program farm equipment to follow. The use of PA has allowed farmers to use less resources, which saves money and reduces environmental impact. However, technology is a two-sided coin, benefiting both end-users, the farmers, and ATPs differently. In agriculture, power asymmetry has been cited as a critical issue existing between farmers and ATPs,and this impacts farmers negatively. For farmers to deploy and have more control over data decision-making on their farms, AI assurance methods need to be integrated into their technologies. There are currently a few studies on this subject in agriculture, but many do not involve agricultural end-users or fall short of meeting the needs of the end-users. If end-users and policymakers are not able to understand how their data is collected and used in the agricultural AI models, they will not be able to make educated decisions about their work. This chapter proposes solutions to benefit all agricultural end-users, including prompting the use of participatory design and adopting more user-centered principles when integrating AI assurance models into agricultural technologies.