Farmers' Behavior and Policy Design in the Era of Smart Farming Technology

The agricultural sector faces a dual challenge: ensuring food security while simultaneously protecting the environment. Population growth, climate change, and environmental degradation exacerbate this challenge. A promising pathway to address it is a shift towards sustainable intensification—that is, achieving higher productivity while reducing negative environmental impact. Smart farming technologies (SFT), particularly those based on artificial intelligence (AI), offer substantial potential to support this transition by enabling autonomous monitoring and site- and time-specific management. Nevertheless, the adoption of these technologies by farmers remains limited, and substantial knowledge gaps persist regarding farmers' behavior towards SFT. At the policy level, the European Union's Common Agricultural Policy aims to promote digitalization and sustainable practices through financial incentives, but such programs have often been criticized as inefficient and ecologically ineffective. SFT could support more results-oriented policy instruments; however, research is lacking on how their capabilities could concretely influence policy design. This dissertation addresses these research gaps through empirical studies that examine the interaction between farmers, SFT, and agricultural policy in Europe. The aim is to deepen the understanding of the factors that influence farmer behavior, how SFT may reshape policy-making, and how optimal policies can leverage the potential of SFT to support sustainable intensification in the agricultural sector. Chapter 2 analyzes how "peer effects"—specifically verbal exchange and field observation among farmers—influence farmers' technology adoption decisions. Using survey data from 313 sugar beet farmers in Germany and a novel, spatially explicit survey tool, we employ a double-selection LASSO approach. The results show that both forms of peer effects significantly affect adoption and mutually reinforce one another. The likelihood of adoption is highest for farmers that observe many fields in close spatial proximity and verbally exchange with many adopters. Chapter 3 investigates farmers' preference for AI-based decision-support tools. Based on an online survey and an embedded economic experiment involving 250 German farmers, the chapter uses a novel Bayesian probabilistic programming approach to quantify the willingness to pay. The findings reveal clear "algorithm aversion": farmers prefer recommendations from human advisors over those generated by AI—even when the AI outperforms the human. The chapter introduces the concept of AI anxiety as a key behavioral factor and proposes its integration in future technology adoption models. Chapter 4 shifts the focus to agricultural policy by examining how smart weeding robots could affect the design of payments for ecosystem services. Using a simulation model, we explore how the robots' capabilities—selective weeding and autonomous monitoring—could enhance the efficiency of both action-based and results-based payments. We find that improved monitoring supports the efficiency of results-based schemes, while selective weeding can improve action-based approaches. Overall, the efficiency of both payment types increases compared to when no robot is used, which shifts the frontier of current policy design options. In sum, this dissertation contributes theoretically, empirically, and methodologically to a better understanding of farmers' behavior towards SFT and identifies how SFT could change agricultural policy design. The findings of this dissertation show that using SFT for sustainable intensification has the potential to make agricultural policies more effective. However, technology introduction alone is not sufficient—appropriate guidance is essential to ensure proper use. Social learning can help to address farmers' algorithm aversion. Policy makers, advisory services and technology developers should work together to facilitate large-scale adoption by clearly communicating benefits and reducing (perceived) efforts for farmers.