Nevertheless, one should be cautious with generalising ResAT findings of a single farming system case. We reason that the CAP’s resilience-enhancing or resilience constraining capabilities are very dependent on the farming system’s characteristics, such as its functions, its regional context and the specific challenges that it faces . For instance, De Veenkoloni¨en faces challenges that are specific to the system , which require specific policy interventions to be able to strengthen the system’s resilience. Farming systems across the EU vary widely in their characteristics and are exposed to different economic, social or environmental stresses and shocks. It is therefore unlikely that the results of De Veenkoloni¨en, with its own specific characteristics and challenges, translate directly to other farming systems. Moreover, Member States vary significantly in their CAP implementation choices, resulting in different goal priorities and configurations of instruments, both in Pillar I and Pillar II. These implementation choices of Member States will determine how the CAP enables or constrains the resilience of farming systems. A logical follow-up study would, therefore, apply the ResAT to multiple different EU farming systems to compare results, leading to a more complete picture of the CAP’s enabling and constraining effects on the resilience of different farming systems. Fourth, the ResAT’s top-down approach appeared to be useful for examining systematically the different extent to which the outputs of public policies are suitable for enabling or constraining the robustness, adaptability, and transformability of complex systems. However, it is important to keep in mind that, if a policy appears to enable resilience, this does not automatically imply that the farming system uses this improved capacity. Therefore, a recommendation for follow-up research is to conduct an in-depth bottom-up case study on how farming system actors experience the influence of policies on the resilience of the system.

Such follow-up research would complement the top-down findings and could help to create more empirical evidence on the relationship between policy outcomes and resilience. Last, the ResAT and its coloured wheels proved to have a discussion initiating character, ebb flow tray which was emphasised by the extensive reflection among the focus group participants on the current way of thinking about resilience and policies. This implies the tool’s usefulness for stimulating discussion with policy practitioners about the resilience effects of public policies. It is important for these discussions to stress that the ResAT does not measure the policy’s actual impact on resilience and that the traffic-light coloured wheels do not imply a normative judgement of the policy. The ResAT should, therefore, always be accompanied with an explanation of the analysis and the results that specifies the purpose of the tool. In data servers rented from ‘cloud’ providers such as Amazon Web Services, and across fibre optical cables connecting financial centres such as Chicago, New York, and London, algorithms owned and designed by high-frequency trading firms chase after and exploit milliseconds-long opportunities to make profits . Involving humans but creating “regions of technical autonomy” where they are absent, high-frequency trading revolves around “machine-machine ecologies” proceeding along “an autonomous trajectory in a temporal regime inaccessible to direct conscious intervention” . With around 80% of financial markets now run by machines , this expanding scene of autonomous action signposts the power and social significance of technical cognitive assemblages interacting with human cognition to form a “planetary cognitive ecology” – an especially helpful concept, I argue, because it highlights the rising significance of widespread but often obscure digital technologies in the production of daily life. Farming is constitutive of this emerging ecology because a proportion of the billions of transactions taking place each day on these financial markets always pertains to agricultural commodities . At a time when current and future values and prices of food crops fluctuate according to the autonomous actions of machines, new efforts to create so-called ‘smart farming’ arrangements amplify the relationship between agriculture and digital life .

Underpinned, at least in part, on the claim that digital worlds populated by smart technologies will produce new economic efficiencies, smart farming creates opportunities for firms to collect and use data to then develop valuable insights. A ‘smart’ tractor, for example, flood and drain tray is developed by firms such as John Deere or Kubota with one eye on the field it will harvest, and another eye on connecting data it might be able to harvest with other ‘reserves’ of data . More broadly, smart farming reflects a drive by agricultural technology providers and agricultural transnational corporations to establish strategic positions within the broader digital economy. A useful case in point here is the recent transformation of Monsanto “from an agricultural biotechnology company into a data-science-driven organization” . Its strategy “entails the mass collection of farm data throughsensors attached to everything from tractors to water sources. All the data is fed through a digital platform set up by a service provider, whose algorithms display conditions on the farm and make specific recommendations.” The firm bought startups such as Precision Farming in 2012 for USD 210 million and Climate Corp in 2013 for USD 930 million, while adapting a software platform from the AI firm Data Robot as it sought to “invest in the technologies to gather and study all the data underlying decision-making on farms.” Today, “Hundreds of models run on the platform to develop innovations for the company’s supply chain, its commercial processes, and, of course, farmers.” Monsanto has therefore retooled its employees and increased its data science team from 200 in 2017 to 500 people in 2020. Against this general backdrop, the following paper presents a critical examination of smart farming by culling insights from recent contributions across a wide range of pertinent literatures. I begin by reviewing literature on smart farming with a view to identifying lessons about extant and emerging innovation processes. I follow other critical analyses in recognizing the centrality of innovation processes in generating smart farming products, services, arrangements, and problematic outcomes . I then use insights from critical human geography scholarship on the significance of understanding topological transformations to move beyond interpretations that identify only a narrow range of smart farming problems, such as a lack of coordination or limited uptake by farmers. Instead, I examine a broader set of challenges produced by smart farming developments. The overriding concern, I argue, is that smart farming unfolds via the production of numerous ‘misconfigured innovations.’ Next, inspired by insights from literature on responsible research and innovation , I probe the stakes of looking beyond the misconfigured innovations of smart farming. I discuss how new technologies might come to play a role in producing emancipatory smart farming. As I will explain, ‘actually existing’ smart-farming developments connect with activities on the technological horizon. I pay particular attention to research on the ‘internet of people’ , which paints a stark new picture of social life generally, and rural life in particular.

The provocative scenarios painted by this line of research should alert rural studies scholars to a new scene computed and calculated according to new conceptualizations of sociality and spatiality. Finally, I refer to some recent efforts to introduce ‘smart’ technologies and practices to India. I use this case as a springboard toward identifying requirements for alternative, emancipatory versions of smart farming.The central consideration is the use of information-intensive digital technologies to enable new practices, such as satellite-guided and even auto-steering farm machinery , automatic milking systems , or automated body condition scoring for livestock . The objective is to use digital technologies to make more efficient use of inputs such as labour, pesticides, or fertilizer and thereby improve the quality or quantity of outputs. Smart farming technologies such as soil water outlook tools or new rice varieties conceivably create competitive advantages that justify investment. The vision, then, is that “farming and food will be transformed into smart webs of connected objects that are context-sensitive and can be identified, sensed and controlled remotely […] resulting in new control mechanisms and new business models” . Yet, although the main zone of action for smart farming is the farm itself, these developments in rural space connect activities to wider ecologies. It is impossible to ignore how new digital actions on the farm overlap with numerous other ‘smart’ moves along the food value chain. In supermarkets such as the UK’s Ocado, for example, investments in robotics yield new data-intensive platforms . In restaurants, McDonald’s has established a Tech Lab to integrate startups such as the voice-based Apprente, which it purchased with a view to improving the drive-thru customer experience ; McDonald’s has also invested in Plexure, a New Zealand firm which “creates personalized offers for individuals” when they enter a McDonald’s drive-thru .Elsewhere, artificial intelligence informs voice recognition services in smartphone apps run by companies such as iFood, the Brazilian food delivery company . Soon, these sorts of firms will integrate “unsupervised predictive learning” to work on data in unanticipated ways and, if society awards it authority, conceivably manage “ever more complex algorithm-centric economies” . Indeed, in some scenarios, AI will have the computing power and intelligence to paint pictures of future worlds, and thereafter work on activating socio-technical relations, nft hydroponic according to its vision of what should work . In acknowledging the overlaps and connections between on-farm smart farming developments and more general digital shifts along the food value chain, my aim is to locate smart farming as a crucial zone of action within the ‘planetary cognitive ecology’ . As such, the developments at issue in this paper warrant examination in the light of critical scholarship on digital life .

One core point of this scholarship is that ‘smart’ life generally – and particular developments such as those pertaining to smart farming – creates new sociotechnical realities that empower ‘big tech’ to syphon off or dispossess diverse digital subjects of value generated by analysing data about their activities . Moreover, when “data is the new cash crop” to be harvested from smart farming arrangements, other effects emerge, such as a tendency to enrol automated infrastructures, often governed by opaque, oppressive , and inherently biased proprietary algorithms searching for and expecting to find predictable patterns across socio-ecological environments. Food producers today are confronting new pressures from the operation of algorithms controlled by tech firms or integrated into the emerging operations of supplier, processing, and retail firms up- and downstream of the farm . There are new players on the scene, such as the British online supermarket Ocado using data to glean novel insights about food systems. In the US, moreover, Amazon is pursuing market share along the food chain. As Mooney notes, it might soon be “setting the standards and trends for food security and nutrition.” Thus, per the analytical direction offered by critical data studies, the emergence of smart technologies in the food system – on farms, of course, but also in supermarkets, restaurants, or other foodscapes – requires a focus on uneven effects, with some stakeholders making clear and significant gains, while others lose out in diverse ways. In this regard, smart farming developments also invite analyses informed by scholarship in critical agrarian studies and its insistence on recognizing power asymmetries within the ‘corporate food regime’ . At one extreme, a small number of large and powerful agri-TNCs exert influence up- and downstream of the farm; at another, close to three billion people remain smallholders and small family farmers occupying vulnerable positions within global value chains. Whenever arguments about the virtues of smart farming emerge in the context of the research and development activities driven by agri-TNCs, scholarship in critical agrarian studies suggests alarm bells should ring. One central concern is the role of smart farming in laying the conditions for ‘land grabs’ . Because it is bound up with a broader push to transform and ‘modernize’ agricultural practices, it emerges with a view to widening the ‘yield gap’ between capitalist and peasant sub-sectors of the food system . Pressures on peasants and other smallholders to leave the land will intensify if smart farming operations in capital-intensive settings increase their yields, potentially while reducing their use of expensive agricultural inputs such as pesticides.