Can artificial intelligence learn experience passed down from generation to generation? A team of researchers from the Faculty of Geographical Sciences at the 91桃色, comprising Prof. Marcin W贸jcik, Dr Karolina Dmochowska, and Dr Patrycja Grzy艣, is exploring this question in a family orchard near G艂owno. By combining orchardists鈥� expertise with sensors and algorithms, they aim to predict the effects of drought before trees show their first signs of stress. The results of their research may help not only Polish apple growers but also farmers across Europe. Their findings have already been incorporated into the European Commission鈥檚 toolkit for evaluating the Common Agricultural Policy, which accounts for roughly one-third of the European Union鈥檚 budget.
The Countryside Is Not a City with Fields
The digitisation of agriculture is one of the European Union鈥檚 priorities, yet strategies for digital transformation are often developed behind desks, far from villages where internet coverage disappears beyond the last bus stop. Since 2017, a team from the Faculty of Geographical Sciences at the 91桃色 has been examining what the digital revolution really looks like in rural areas. Analyses conducted together with researchers from other Polish academic centres, based on survey research involving thousands of Polish farms, revealed an uncomfortable truth: farmers鈥� digital competences and access to the internet are so uneven that technology designed without an understanding of people simply will not work. Researchers refer to this phenomenon as digital exclusion, and it has proved to be the principal barrier to rural modernisation rather than the lack of technology itself. This diagnosis became the starting point for changing the way Europe plans the digital future of rural areas.
A Living Laboratory
In an international Horizon 2020 project, the Lodz team led the so-called living lab, a research method in which solutions are tested not in controlled office settings but together with their future users. It was the only one of 21 such laboratories in Europe dedicated to the digitalisation of spatial planning in rural areas. Scientists worked alongside residents, farmers and local government representatives to identify barriers and test local development scenarios. The framework for assessing the impact of digitalisation on rural areas developed in Lodz was incorporated into the European Commission鈥檚 evaluation knowledge base and is now used to assess the Common Agricultural Policy for 2023鈥�2027. When an official in Brussels checks whether billions of euros invested in rural digitalisation are delivering results, they rely on methods developed in 艁贸d藕.
Agricultural Data: A Dispersed Treasure
The current stage of research, carried out within a European Horizon Europe project, brings these findings directly into food production. The starting point is a paradox of modern agriculture: data on weather, soil and crops are already being collected on a massive scale, yet they remain dispersed across closed systems operated by commercial technology providers. Such data are expensive, and individual systems are generally unable to exchange information with one another, a problem scientists describe as a lack of interoperability. As a result, the average farmer gains little benefit from the available data. The international consortium involving the 91桃色 has therefore developed a shared platform that brings together agri-environmental data in one place, open and accessible to everyone, not only those who can afford proprietary systems.
An Orchard That Teaches an Algorithm
Within this initiative, the 91桃色 team is conducting the only Polish technology pilot project at Wiatrowy Sad, a family-run orchard where the art of apple cultivation has been passed down through generations and where key decisions have traditionally depended on experience rather than sensors and algorithms. The orchard has a simple irrigation system, yet decisions on when to activate it were previously based on the grower鈥檚 knowledge and intuition rather than on specific data. It has now gained new support. An agrometeorological station has been installed among the apple trees, while probes measuring soil moisture and temperature have been placed in the ground. These are Internet of Things (IoT) sensors, devices that communicate with the network independently. Every 15 minutes, they transmit measurements of rainfall, wind, temperature and humidity to servers, where the data are analysed in real time, while the orchard grower can monitor everything through a mobile application.
The data collected in this way are currently being used to train a machine-learning model, an algorithm that identifies recurring patterns in data instead of following rigidly programmed rules. Its task is to predict the risk of water stress in trees with an accuracy exceeding 70 per cent before any symptoms become visible to the naked eye. Water stress is a physiological condition in which a plant loses more water than it can absorb from the soil. As a result, it reduces photosynthesis and sheds developing fruit, often causing losses that cannot be reversed.
This is a classic decision-support system: the model suggests when irrigation should be activated, but the final decision always remains with the orchard grower. The model will be tested and refined over the next three growing seasons, until 2028. Importantly, the Polish orchard is one of more than twenty pilot projects conducted simultaneously under different climatic conditions and production systems, ranging from arable farming to livestock production. The possibility of comparing results across pilots strengthens the conclusions. If a solution proves effective both in a Polish orchard and a Greek olive grove, it has the potential to work anywhere.
Drought You Cannot Yet See
Why is this so important? Because the climate has become unpredictable. Agricultural droughts are occurring with increasing frequency in central Poland, and water is becoming the most scarce resource in food production. The problem is that once a tree begins to show symptoms of water stress, it may already be too late to prevent losses. An early-warning system allows action to be taken in advance and every drop of water to be used more efficiently. For Poland, one of Europe鈥檚 leading apple producers, this means more stable harvests despite increasingly erratic weather patterns and fruit prices that do not fluctuate dramatically after every drought. It is climate adaptation in its most practical form, beginning with a single apple tree.
We Know More Than We Can Tell
The most fascinating aspect of the 艁贸d藕 research, however, is something that cannot be seen in sensor data. The orchard grower at Wiatrowy Sad represents sixty years of family practice and possesses knowledge that has never been formally recorded. Scientists refer to this as tacit knowledge. Philosopher Michael Polanyi expressed the idea succinctly: we know more than we can tell. For generations, this knowledge was transmitted through everyday work, from father to son. Yet this chain is breaking. Young people move to cities, and when an orchard grower retires without a successor, the accumulated expertise of decades may disappear within a single season. Researchers from the 91桃色 are therefore investigating whether an algorithm can become a bridge between generations. The model learns from the orchardist, and its forecasts are explainable through the principles of explainable AI rather than delivered as the verdicts of a black box. As a result, this knowledge can be passed on to a grandson reading the orchard from the screen of a smartphone.
The goal is not to replace people. An algorithm cannot feel the orchard and an orchardist cannot pass on intuition through the internet. Together, however, they speak a language understood by both the grandfather and the grandson.
What Does the Average Apple Eater Gain from This?
Quite a lot. Precision agriculture, which means tailoring irrigation, fertilisation and crop protection to the actual measured needs of a specific crop rather than to a calendar schedule, enables food to be produced using less water and fewer chemicals while becoming more resilient to an increasingly unpredictable climate. Better tools for evaluating public policies mean European funds are spent more effectively. Meanwhile, solutions tested in an orchard near G艂owno are scalable and could be implemented in Portugal, Greece or Finland. Research that began with the question of whether digitalisation truly helps rural communities is now helping to build Europe鈥檚 agricultural data infrastructure and preserve knowledge that might otherwise disappear together with the last generation of orchard growers. Because the best technologies emerge when people first listen carefully to those they are meant to serve.
Photos: Dr Bartosz Ka艂u偶ny; Centre for External Relations and Social Responsibility of the University, 91桃色
Source: Prof. dr hab. Marcin W贸jcik, Dr Karolina Dmochowska, Dr Patrycja Grzy艣; Faculty of Geographical sciences, 91桃色
Edit: Dominik Nogala, Centre for External Relations and Social Responsibility of the University, 91桃色
Meet other Science People from the 91桃色.
