Why AI has not yet had its breakthrough in agriculture
Artificial intelligence is gaining ground in many areas. However, the new technology does not yet seem to have really arrived in agriculture. The reason for this is nature, which is throwing a spanner in the works of AI. Nevertheless, the opportunities that AI could offer agriculture are immense.
Wednesday, May 1, 2024
Artificial intelligence has been the talk of the town ever since ChatGPT was founded. While the technology is already being used in numerous areas and is already making processes more efficient, it is still unclear how AI can be used in other areas. One of these is agriculture.
While digitalisation has arrived in agriculture, artificial intelligence still needs to be trained to a large extent in order to be used efficiently in agriculture. The biggest difficulty here is that nature is often unpredictable. ‘From the point of view of AI specialists, there is nothing more brutal than nature,’ explains Philipp Schmid, Head of Research at the Centre Suisse d'Electronique et de Microtechnique (CSEM), in an interview with the ‘Bauern Zeitung’. According to Thomas Anken from the Digital Production research group at Agroscope, the problem often lies in the collection of data. AI is dependent on good data, but this is often in short supply in agriculture. ‘There are still a lot of gaps in the area of sensor systems, as many important parameters cannot yet be collected easily and cost-effectively.’ One example of this is the nitrate content in the soil.
Measuring biodiversity with a super drone
Another hurdle for AI in agriculture is the small areas of land. As Schmid explains, Switzerland is full of small farms with different farming zones and many special circumstances - poor conditions for affordable AI applications.
However, according to Schmid, farmers are generally not averse to new technologies or AI. The rapid spread of autonomous barn robots or drones also makes this clear. Just recently, a new drone that can measure biodiversity in forests made the headlines. According to a report in ‘Schweizer Bauer’, the drone was developed by a team of researchers from the Swiss Federal Institute for Forest, Snow and Landscape Research WSL, ETH Zurich and the University of Pisa. Unlike conventional drones, the monitoring drone developed by the scientists does not perceive shrubs and branches as an obstacle, but can push them away or even glide through them and thus observe biodiversity.
When laypeople measure biodiversity
The Flora-Incognita app offers another way of monitoring biodiversity. It makes it possible to identify plants quickly and easily. Using the app is child's play: users take a photo of the plant they want to find out more about with their smartphone camera, upload it to the app and then receive information about the plant. This information is provided by botanists.
The best thing is that biodiversity is monitored entirely by means of citizen science. Each individual user plant photo, together with its location and time, contributes to a large data set that ultimately provides information on biodiversity in the individual areas.
These are just a few examples. One thing is clear: AI offers many opportunities for agriculture. The new technologies can not only improve efficiency, productivity and sustainability, but also increase animal welfare and reduce the physical strain and time commitment of farmers.
Nevertheless, AI also harbours certain risks. For example, cybercrime and data misuse are always an issue. According to Anken, it is therefore important to take the necessary precautions to prevent such misuse.
Kindly note:
We, a non-native editorial team value clear and faultless communication. At times we have to prioritize speed over perfection, utilizing tools, that are still learning.
We are deepL sorry for any observed stylistic or spelling errors.
Related articles
Production: More food with fewer resources
One of the greatest challenges of this century will be to feed an estimated ten billion people in a safe and sustainable manner. To achieve this goal, agriculture needs to become much more productive.
From Data to Harvests - How Digitization is Improving Agriculture
Digitalization is making its way into agriculture. At the Swiss-Food Talk on April 25, 2023, three experts from the agricultural machinery industry, vegetable production, and agricultural media discussed how digitization is changing food production. The consensus is that we are in the transition from industrial to smart agriculture, where data and algorithms as support allow precise interventions and serve sustainability.
Green genetic engineering: A rethink is required
There are now more skeptics than ever before when it comes to biotechnological plant breeding methods, despite 30 years of research having produced a clear data basis. Conventional genetic engineering or the more modern CRISPR/Cas method present no increased risks compared to traditional breeding methods, such as cross-breeding.
Residues, thresholds, trust – looking calmly behind the headlines
In this conversation with toxicologist Lothar Aicher, the discussion focuses on how residues enter the body, how their potential harm is assessed, and what role modern analytical methods play.
Hazard is not the same as risk: how we understand – and should understand – threshold values
In this episode of the podcast, risk researcher Angela Bearth discusses residues and threshold values in food – a topic often debated emotionally. In public discussions, threshold values tend to dominate, while the complex challenges of agriculture are rarely considered. Studies show, however, that once these relationships are explained and solutions are addressed, people respond in a more differentiated way.
PFAS, trade-offs and responsibility – how politics and agriculture find solutions
In this episode of the joint series by Agrarpolitik – der Podcast and swiss-food.ch, National Councillor Christine Badertscher discusses how residues and threshold values are debated in Parliament.
Thresholds, approvals, responsibility – how plant protection products are really assessed
Threshold values are often at the centre of public debate – yet in reality they are only a small part of a much broader risk assessment system. Dr Michael Beer, Head of the Food and Nutrition Division at the Federal Food Safety Office, provides clarity.