Like all economic sectors, agriculture is experiencing the rapid introduction of digital technologies. Since the mid-2010s, the concept of "digital agriculture" has emerged. What does it mean?

Inrae (the French National Research Institute for Agriculture, Food and the Environment) and Inria (the French National Institute for Research in Digital Science and Technology) have jointly undertaken, in a White PaperThe aim is to better understand, control, prepare, equip and support the deployment of digital technology in agriculture and the food chain, taking into account the way it will transform sectors and their ecosystems, with the aim of putting it at the service of the agro-ecological transition (AET), the territorialization of food and rebalanced supply chains.

This White Paper details the challenges of transforming agriculture and food systems, reviews existing digital technologies, examines the possibilities offered by digital technology for the agro-ecological transition, and identifies the risks associated with the uncontrolled development of digital agriculture. It concludes with technical issues and challenges for developing responsible digital agriculture.

Référence : 1

Digital and agriculture: towards agroecology?

The production, therefore, could evolve towards two models:

• sustainable intensification (improvement of process efficiency and insertion in long supply chains);
• Agroecology, which relies on natural processes to produce and which is part of local and sovereign food systems. It is abouta set of practices that aim to improve farming systems by "mimicking" the natural processes, thus creating beneficial biological interactions and synergies between the components of the agroecosystem.

It is precisely this path of " digital technology at the service of the transition to agroecology and the renewal of food systems " that they explore in the six chapters of this white paper.

The levers of digital agriculture

• Data: " Sensors, the source of data acquired in the field, pose hardware and software challenges: the nature of the quantity to be measured, the preferred measurement technology(ies) and the way to implement it (them) to obtain useful information at a lower cost must be defined. Free satellite images, connected objects and collaborative applications on cell phones are massive sources of data.
• Modeling: " This is the key element in the representation of agro-ecosystems, which are complex by nature, in order to simulate, optimize and control them. The scales range from the plant or animal to the population, the territory or the value chain, with the added challenge of coupling the scales and models representing the subsystems. Modeling, an old approach in agronomy, is now being renewed thanks to digital technology.
• Simulation: it is used "to represent agrosystems, or even socio-ecosystems, with behaviors that are difficult to analyze; it provides a description of possible states and has many uses, such as individual or collective decision support (support models), training, etc.". Optimization, on the other hand, goes further in decision support: "it seeks solutions to a given problem according to one or more criteria".
• Knowledge extraction: " In addition to these analytical modeling approaches, new families are emerging with models directly inferred from data, when the latter are sufficiently numerous to cover the space of possibilities. This is the case of remote sensing data or time series collected via connected objects). "After a necessary pre-processing to improve the reliability of these data ("cleaning", reconciliation with expert data), they are processed via different formalisms to extract intelligible information. The new knowledge that is extracted or generated by the models is formalized and organized "in order to be delivered to various audiences, via decision support tools adapted to each activity, whether it is related to culture or breeding, and at all scales.
• Robotics: automated and/or robotized systems are increasingly precise and reliable. " If robotics was initially developed in the livestock industry (milking robots with fixed stations, cleaning robots in closed environments), it faces additional challenges in crop production (uncontrolled, changing external environment...) .

Digital opportunities for agroecology and sustainable food

The concept of precision agriculture or breeding is today more associated with intensive agriculture: "it is nonetheless valid in agro-ecology, in particular to monitor the good health of plants and animals, based on automated observations via sensors and models, but also to implement - on a large scale - more complex cultivation processes (crop associations, selective collection...)".

This requires sensors and models capable of analyzing the received signal to give a description of the state, a prediction of the future state or a prescription.

Digital technology could enable the renewal of the agricultural ecosystem, including agricultural services (insurance, consulting), the organization of value chains and the management of agricultural territories. "Value chains are transformed by disintermediation, favored by the Internet, but also by the possibility of "transparency" on the history of the product, which is now increasingly demanded by consumers."

Another facet of the agricultural ecosystem, the management of territories is impacted by digital technology. " The scale of territories is relevant in agroecology (landscape ecology, looping of cycles via the circular economy) and agriculture, which has a central place in territories, is the object of tensions related to the use of resources (land, water) or its role in ecological services.

Finally, digital technology could provide tools to better identify material flows and facilitate mediation and collective decision-making.

Because of connectivity, digital sciences and technologies facilitate individual and collective sharing and learning, which are sources of innovation in agro-ecology. "Knowledge (including traditional knowledge) is capitalized and exchanged between peers, either directly (social networks) or in participatory collective processes that increasingly integrate digital technology.

Finally, " the farmer could become a supplier of data to private or public actors (research via on-farm experimentation, territorial documentation, etc.), which could change his status in the direction of better integration and recognition.

Most international organizations and donors see digital technology as a major source of transformation in the countries of the South and in particular in Africa. "Digital technology could help diversify the service economy, accelerate the structural transformation of agriculture and make it more attractive to young people, improve local value chains (building territorialized food systems) or long-distance supply chains (guaranteeing product traceability), and help build the information capital of territories.

The risks of increasing technologization

The first risk identified is that of disappointing the expectations of a more ecological agriculture. "If the development of digital technology in agriculture provides solutions for reducing inputs, this gain could be accompanied by a technological lock-in that could hinder the implementation of more radical and more systemic alternative practices and organizations that could lead to greater environmental and socioeconomic gains.

The widespread implementation of digital interfaces between the farmer and the animals or plants, in a process of increasing technologization of agricultural production, also runs the risk of "bringing about a kind of loss of link to nature (and in particular in the human-animal link), whereas society is undoubtedly expecting an agriculture with a stronger link to the living world around us.

The ecological footprint of digital technology is not yet well known in agriculture: "the multiplication of equipment and operations for data capture, transfer, storage and calculation could degrade the environmental balance of digital agriculture".

Another family of risks lies in the social consequences of the reinforcement, through digital technology, of a trajectory of industrialization, with a concentration of production in ever larger, productivity-oriented units and farms. "This movement would lead to risks of exclusion of minority forms of agriculture, mainly for farms of small economic size. The development of robotics could reinforce the precariousness of agricultural work, especially for poor immigrant workers. Difficulties in accessing digital technologies would also be a factor of exclusion in agriculture, whether on an individual level (lack of skills) or on a territorial level (lack of digital infrastructure).

Digitization could also have consequences on farmers' decision-making autonomy, and even on the meaning they give to their profession, with their fear of becoming mere " data workers".

Third family of risks: digital and food sovereignty. " The increasing digitization of the food chain may lead to an integration of agriculture, with the appearance of monopolistic actors and tools."

Digital sovereignty also involves control of data: the authors of the White Paper point out " a risk of confiscation of agricultural data by suppliers of digital technologies or services (agricultural equipment, AgTech companies, digital giants)".

Cybersecurity risks also need to be taken into account: " Relatively unaffected today, our food systems are vitally important, which could turn them into potential targets in the future.

Finally, the digitization of the agri-food system risks increasing dependencies between the different actors of this system and creating new ones with the actors who will produce and own these technologies. "This risks accentuating the vulnerabilities of this system in the face of the many shocks that will inevitably and heavily affect the functioning of our societies in the coming decades (...). The development of digital technology can amplify the dynamics of complexity, but we must avoid a technological headlong rush that would lock us into a spiral of uncontrolled complexity.

In conclusion, the authors of the White Paper detail the scientific and technical challenges as well as the associated human challenges:

• better manage collectively, by integrating the scale of the territories;
• better manage the farm ;
• rebalance the value chain, from upstream to downstream;
• create and share data and knowledge.

Inria White Papers