Abstract
Farming and emerging technologies, including Artificial Intelligence (AI), have been promoted as being key interventions for addressing both global food shortages and adverse impacts of climate change. Smart technologies and AI are increasingly touted for their ability to improve productivity and increase efficiency through better farm-level decision-making and resource utilization. This emerging global discourse has induced great interest and investment in AI and robotics development and adoption in farming sectors around the world. Despite the growing interest and investment in farming technology there is limited research that is investigating the sustainability of AI technologies – particularly given some scholars’ highlighting of potential socio-ecological risks and costs that may come with AI technology. Where academic literature exists, it focuses on sustainability relating to the technology as an artefact in itself, as opposed to consideration of the spaces where the technology is perceived, funded, designed, developed, and lived. The research presented in this thesis builds on emerging work on sustainability of AI, by exploring how sustainability of AI is imagined in situated socio-material co-design spaces, and how insights from different actors and spaces might shape the possibilities of sustainable AI in co-design spaces and the wider Agtech ecosystem.
The research project was situated within a social research team, within a large Agtech project in Aotearoa New Zealand, alongside a group of parallel social scientific studies looking at different socio-ecological dynamics relating to design and adoption of Agtech in Aotearoa New Zealand. This research reported in this thesis adopted a transdisciplinary qualitative approach to research design – informed by Feminist STS theoretical framing and discussions about the importance of socio-technical imaginaries - by exploring in-depth insights from a transdisciplinary group of participants drawn from the wider Agtech ecosystem in Aotearoa New Zealand. Twenty-nine participants were iteratively broken out into three clusters of participants (involved in: farm spaces, technology design spaces and policy spaces), and interviewed about their interest in sustainability, interaction, and experience with AI Agtech, and their imaginaries about sustainability of AI. Qualitative insights were collected through semi-structured interviews with key participants: a group that included farmers, people from the wine-grape, apple, and blue-berry industries, farm managers, technology developers, agronomists, social science scholars, and government policy advisors.
The results of these interviews show that within co-design spaces, there are different and distinct socio-material spaces, with embedded norms and practices that shape how sustainability of AI imaginaries are constructed in these spaces, including how actual practices of sustainability are enacted. First, the participants in the farm cluster revealed farms as socio-materially complex spaces which are guided by economic and socio-ecological values. Farm spaces involved participants who were not only aware of the complex relationships between their social and material parts of the farms, but also allowed themselves to shape and be shaped by these co-evolving and active relationships towards shaping their farms spaces. These socio-material dynamics shaped how farmers and farm-related industry participants viewed their farms, perceived power, assembled their farms, and imagined sustainability of AI within farm spaces. In the second cluster of participants, labs were less complex, they were replicable and reproducible – mainly guided by technical, client and economic values. This influenced how technology developers imagined sustainability within technology design spaces, and in the adoption spaces. Finally, policy spaces were more nebulous, dominantly trapped within global discourses about sustainability and AI, and mainly guided by economic value, national values, and narratives of technological growth. The results also opened-up opportunities and possibilities for a more inclusive dialogue of different perspectives towards shaping sustainability of AI.
In conclusion, sustainability of AI remains a complex concept that can only be understood by focusing on the situatedness and the social-material relations within different spaces and in specific situated co-design processes. This thesis advocates for the creation and transformation of co-design spaces, into spaces where power sensitive conversations are not only embraced, but different actors are willing to move from their comfortable spaces to other socio-material spaces - with the aim of building their awareness and understanding the different dynamics that shape those spaces. As such, technology designers and those who fund technology, should invest in design and adoption technology that articulates the values and the needs of the end-users and their environments.