Conducting agroecological research in the field demands choosing or modifying natural places to tailor them to machines and quantitative measurements, ensuring the production of reliable, consistent data while navigating the myriad challenges inherent in unpredictable environments where unexpected occurrences are commonplace (Kohler, 2002). Preparing the field for AI and data intensive agroecological research involves the meticulous construction of objects that can be investigated (cf. Wylie (2015) on fossil construction). These encompass a diverse array, from fruits and insects to data and technologies. Although not always acknowledged as scientific labor, such activity plays a pivotal role in laying the foundation for meaningful research outcomes. When examining pest-plant interactions (Giron et al., 2018; Parker and Gilbert, 2004), the complexity intensifies because it involves the dynamic interplay between different living organisms, each with distinct physiological needs essential for survival and biological constraints that limit the extent to which they can be altered. This symposium, organized as part of the Phil_Os project (https://opensciencestudies.eu), closely investigates the ways in which field and objects are prepared for AI and data intensive agroecological research, with a specific focus on the construction of plants, pests, and their interaction. To these aims, the first presentation by Cavazzoni explores three key dimensions that significantly influence the process: social relations, the environment, and the methods employed. The demands of both plants and pests, along with the added complexities of their interactions, pose unique challenges when preparing the field for AI and data intensive research. Giannetti et al. give an overview of the current state of new technologies used for monitoring and studying insects in both natural and agroecosystem environments. Finally, Leonelli and Ankeny investigate how organisms themselves can be developed as technologies, exploring practices associated with their use both as material tools and as representations.
This presentation closely investigates the ways in which field and objects are prepared for AI and data intensive agroecological research. The discussion is centered around three key dimensions that significantly influence the process. The first one pertains to the intricate tapestry of social relations. This includes how the division and integration of labor and expertise, along with the resulting dynamics, shape the direction of object construction and field preparation. The second axe revolves around the environment. The preparation of the field for automated agroecological research is shaped by factors such as unpredictable weather patterns and the lack of model species due to high biodiversity. Being concerned with pests and plants in natural fields rather than in controlled lab environments results in researchers having limited control over parameters such as temperature, humidity, and light exposure (De Bont, 2015; Knorr-Cetina, 1992; Kohler, 2002). The third crucial dimension to consider is the methods employed. Decisions regarding which aspects to monitor and how to integrate technologies with field elements such as territory, species composition, ecology, and climate greatly influence the preparation of the field and the construction of the objects involved. To effectively address challenges in agricultural development and food security, it is fundamental to align the construction of pests and plants to technological advancements that are not only technically achievable but also useful given pest control methods already on the ground. I ground my reflections on six months of ethnographic work and collaborations with Haly.Id, a Horizon project based in Northern Italy which deals with a plethora of objects such as data, insects, and fruits, and develops innovative technologies for a targeted monitoring of the presence in crop fields of the highly invasive pest Halyomorpha Halys (H. halys) (Ferrari et al., 2023; Giannetti et al., 2024).
The increase of technology advances like drones, rovers, trail cameras, and acoustic sensors that integrate artificial intelligence, along with their higher accessibility, hold the potential to revolutionize the study of animal ecology. Automated monitoring and species recognition in both natural habitats and agroecosystems already represent an important tool in biodiversity management and conservation. However, the application of these technologies for invertebrate monitoring remains relatively limited. The expansion and optimization of such tools would allow for more efficient use of operators' time by reducing sample management and determination times, offering a greater temporal resolution compared to traditional monitoring systems, and facilitating and speeding up necessary control and management actions. This overview aims to present the current state of new technologies used for monitoring and studying insects in both natural and agroecosystem environments, highlighting their challenges, criticisms and potential for future development.
This paper explores an emerging set of scientific practices associated with the use of model organisms in toxicological bioremediation both as tools and as representations. Organisms such as Daphnia are developed and used as technologies to monitor and assess chemicals in polluted waters, and hence serve as sentinel species and diagnostic agents. They also can be used for bioremediation for instance to reduce hazards in chemical mixtures contained in waterways in the environment. At the same time, organisms thus instrumentalised are studied to assess which strains are most well-adapted to chemical pollution – with the ultimate goal to develop models for reduced chemical sensitivity that can be projected onto and investigated in other organisms, notably humans. Hence in this domain, these organisms bridge the field and the lab through the simultaneous use of novel data-intensive approaches appropriate in complex real-world settings and of traditional practices and understandings associated with model organisms from the lab. We argue that such organisms can be viewed as material technologies that nevertheless retain their representational power, and provide an important example of a type of hybridity that will be increasingly common as field-based research leverages the knowledge, data, and technologies created in lab settings.