Written by Andrés García Higuera.
Artificial intelligence (AI) has been the subject of such strong political and social debate that the question of its suitability for its main original purpose – improving supply chain efficiency – may come as a surprise. What if AI really could help strategic sectors cope with pressure? More specifically, could the agri-food sector benefit from this technology to compensate for the shortages broadly forecast as a result of today’s crises?
The crisis resulting from the unprovoked Russian attack on Ukraine shows the huge impact that disruptions on supply chains can have on the EU economy, not least in its agri-food sector. Providing outstanding benefits and challenges, AI has applications at various stages of the agri-food chain, including but not limited to precision farming, value chain integrity, personalised nutrition, reduction and prevention of food waste, enhancement of food safety, and transparency and traceability in the agri-food chain. The technology also has the potential to reshape the agri-food sector; mark a successful transition to climate-neutral Agriculture 4.0; and spur progress toward meeting Sustainable Development Goal 2 to end hunger, achieve food security, improve nutrition and promote sustainable agriculture.
According to a recent Eurobarometer publication on agriculture, nearly half of Europeans think that securing a stable food supply in the EU at all times should be a main common agricultural policy (CAP) objective. As with other technological advances, AI in this domain comes with its own set of benefits, risks, ethical issues and societal implications. Questions raised with respect to AI include how to balance potential benefits against possible risks; how to govern the use of these technologies; and how to incorporate socio-ethical value considerations into the policy and legal frameworks under development.
The reinforcing effect of combining AI with other new technologies is sure to disrupt many sectors, including agri-food. As the role of AI in agriculture and food production has grown in importance in recent years, the agricultural sector has witnessed increased use of sophisticated equipment, such as robots, satellites, drones, other automated vehicles and sensor-based monitoring and irrigation systems. These pieces of machinery serve as sources of data, for example concerning production processes and conditions on the farm, including data on crop growth, soil characteristics, pests, and weather conditions. In automated systems, AI allows real-time monitoring and analysis of agricultural processes, generating critical knowledge to fine-tune strategies for optimal resource utilisation, boosting productivity while minimising environmental impact. The use of AI in supply chain management is also gaining relevance, as the provision of seeds, fertilisers, cattle feed and water resources becomes more complex; and the same is happening with the final products. Tracking systems based on the internet of things (IoT), such as those relying on radio-frequency identification (RFID), have become essential to improving supply chain efficiency and ensuring product quality and customer safety.
Potential impacts and developments
A recent STOA study analysed the potential of precision agriculture for the future of farming in Europe. The potential applications of AI in the agri-food sector go even further. They can help to improve the way food is processed, packaged, stored, transported, prepared, served, eaten and not wasted. Along the agri-food chain, automation, robotisation and AI can help achieve greater productivity while reducing the need for a human workforce. In addition, plant breeding can help to make food production more sustainable, by developing crop varieties that require fewer inputs, for instance. Moreover, genome editing can enable the targeted alteration of a few DNA letters within the existing genetic blueprint of an organism. From the identification of genetic variants to agricultural production and changes in consumer behaviour, AI has the potential to become a game-changing technology in the agri-food domain.
While holding great promise, this rapidly developing field nevertheless raises concerns regarding equitable access, privacy and liability, as well as bias, inclusiveness, accuracy, data set availability and representativeness, data ownership, cybersecurity and the terms used to integrate big data and AI technologies into agri-farm systems. As various AI applications are gradually implemented, they may lead to the loss of user self-determination and agency, and widen the digital divide. They may also open the agri-food sector to non-traditional actors, support new generations of farmers, contribute to the loss of traditional jobs, and enable the development of new business models. Lastly, the digitalisation of agriculture, including AI applications, does not automatically lead to greater sustainability. If data and AI-based solutions are geared only to making traditional agricultural practices more productive, this could actually amplify the negative impact of unsustainable farming practices. It is essential to use AI’s potential to address the profound challenges facing the current food system and ensure citizens enjoy the benefits of AI without being exposed to unnecessary risks.
The application of AI in the agri-food chain has to be considered in the context of the objectives and initiatives relating to the relevant EU legislation and policies. These include the European Green Deal, the CAP, the farm to fork strategy, as well as the proposed AI act. Attention should also be paid to the agricultural specificities of different EU regions and Member States, and the diversity of actors in the agri-food chain.
The proposed EU AI act’s high-risk AI list (Annex III) does not explicitly mention AI applications in agriculture. However, it could be argued that several actual or foreseeable AI applications in agriculture would fall within the scope of that list, especially since that list is neither exhaustive nor fixed. As AI can control food production, even select livestock or which crops to grow depending on a series of factors that may include geopolitical and strategical considerations, this would mean increasing the collection and sharing of data – to be used first to adjust decision-making processes based on machine learning and then as inputs for subsequent decisions. To make these strategical decisions as efficient as possible, farmers will have to share data to an extent that may present problems in a competitive market. And this may include the collectivisation of data at supranational level, which may also pose geo-strategical concerns. These considerations need to be taken into account while developing European legislation, such as the data act still under discussion, and also when deploying adopted regulations, such as the Data Governance Act.
The growing complexity of the management process in the agri-food sector can also have the effect of increasing inequalities by making these tools more readily available to bigger producers and leaving smaller farmers, unable to adapt, behind. Furthermore, there is growing concern about bias in the decision-making process that may prioritise some producers over others, depending on their size, type of production or region. Ensuring proper deployment of AI applications would require the development of a risk governance framework for anticipating and predicting concerns relating to data management and AI applications in the agri-food domain. This would have to include the classification and assessment of risks, as well as risk monitoring and management. A strict liability approach can be complemented with legislative sandboxes, given the need to maintain a balance between the objective of protecting people and society on the one hand and fostering innovation arising from the constant and rapid developments in the AI field on the other.
There is much to be gained, given that AI was first developed by industry as a tool to help improve efficiency in production and logistics. Crisis management requires the same elements as coping with the complexity and unpredictability of supply chains. In the context of the European Green Deal, AI has huge potential to help optimise the management of production and distribution of strategic goods – applications very close to its original purpose. These could include: microchips; water and energy generation and transport (from hydrocarbons to renewables, hydrogen and grid management); as well as fertilisers; pesticides; and food products such as meat and grain.
Read this ‘at a glance’ on ‘What if AI could make the agri-food sector more resilient?‘ in the Think Tank pages of the European Parliament.
Listen to podcast ‘What if AI could make the agri-food sector more resilient?’ on YouTube.