The panel discussion on the DLG Expert Stage raised a question that is central to the future of swine farming: How can artificial intelligence contribute to animal health and welfare, from the barn to the slaughter hook? The panel discussion featured Dr. Danielle Rhen from SLA Software Logistik Artland GmbH, Marielena Krieger from the Goldschmaus Group, where she serves as animal welfare officer, and Mirjam Lechner, an agricultural consultant at Stall Aktiv. Prof. Dr. Heinz Bernhardt from the Chair of Agricultural Systems Engineering at the Technical University of Munich moderated the discussion.
AI Has Long Since Become Mainstream
Mirjam Lechner opened the discussion with an observation that served as a common thread throughout the conversation: AI has long since become a standard feature in every smartphone and in business analytics. Nevertheless, it is sometimes viewed critically in the agricultural sector. Lechner, who studied the meat industry and actively advocates for improved, animal-welfare-oriented pig farming, reported that in her field of work, there is hardly any area that can function without AI. She explicitly views AI systems as assistance systems. Today, monitoring systems are available, but care robots are not yet a reality. Empathetic people remain indispensable in the barn and must be trained accordingly. “A pig is not a screw,” she emphasized. Anyone walking through the barns touches the animals, and this closeness must not be lost—not least to avoid jeopardizing society’s acceptance of livestock farming.
Dust, Ammonia, and Moisture as the Biggest Obstacles
The panel identified dust, ammonia, and moisture as the technology’s main adversaries. These factors, they noted, pose the greatest technical challenge for digital systems operating directly in the barn. Dr. Danielle Rhen, a physicist by training who oversees AI projects in Quakenbrück, made it clear that data processing essentially involves preparing, storing, and forwarding information to the right places. It is crucial, she said, that a camera delivers the same quality throughout its entire operational life: “Because only a reliable database, updated daily, makes the analysis robust.” From an animal welfare perspective, she added, the truth hangs on the hook. That is where it becomes clear how the animals actually fared on the farm.
First the Goal, Then the Data
It was emphasized repeatedly that data quality is not something that emerges only in hindsight. Anyone who wants to ensure that the data is accurate must clearly define the goal in advance and work closely with those in the field to achieve it. Lechner illustrated this with a project to measure tail lengths: “I stood there with a folding ruler and measured the animals’ tail lengths—they ranged from 20 to 40 cm. The AI, which was running in parallel, interpreted a 40-cm tail as 19 cm.” If the measured values are incorrect, the AI simply cannot make sense of them. Furthermore, the AI must be capable of making predictions. She gave the following example: At what temperature must a piglet of a certain size consume how much drinking water, and what happens if it doesn’t? Such questions arise much earlier than the subsequent collection of data.
Lechner also reported that indicators have already been established in AI systems for the lungs that signal stress. For these systems, it makes no difference whether a pig comes from a conventional or an organic farm. Another approach involves the animals’ vocalizations. Lechner explained that sounds indicating pain or illness can be distinguished from other noises. Healthy pigs are also significantly louder, while sick animals become quieter. If this shift can be reliably detected, it would be a major benefit for animal protection and animal welfare.
Who owns the data?
The issue of data ownership and security is a concern for many farmers, as several comments during the discussion showed. Marielena Krieger clarified that the data belongs to the farmer, but also to the slaughterhouse, since it is collected in the slaughterhouse’s system. She emphasized that it is crucial for farmers and agricultural advisors to receive a clear interpretation of this data so they can draw conclusions from it. The panel agreed that the data is generally representative, meaning that one can rely on facts rather than assumptions. Nevertheless, the data must be understood, and it cannot be passed on to farmers unfiltered. The security of data transmission is also a key requirement in this context.
Krieger cited an example in which healed injuries on a pig’s ear allow conclusions to be drawn about upstream problems on the farm. It is precisely at this point that farmers and advisors can be guided. The Goldschmaus Group is working accordingly with AI at the slaughterhouse and using the data obtained to improve processes in the finishing operations step by step.
Technology Must Be User-Friendly
The phrase “taking users by the hand” led to the question of how such support can be technically implemented in a way that actually works in everyday life. The panel agreed that systems will only be accepted if they are understandable, reliable, and integrated into farm operations. This also includes actively planning for emergencies to ensure that the animals are cared for at all times. Prof. Bernhardt added from a research perspective that caution is needed when conducting trials in barns. Anything tested on a larger scale is quickly classified as animal testing, which requires approval.
Another aspect was the next generation. Rhen explained that her company wants to rekindle young people’s enthusiasm for careers in slaughterhouses and is committed to providing support to that end.
Outlook: What We Want to See at EuroTier 2028
In closing, Prof. Bernhardt asked what innovations the participants expected to see at EuroTier 2028. Mirjam Lechner articulated a clear vision: no more nipple drinkers for suckling piglets. Marielena Krieger countered that humans could never be replaced and that AI was not a solution to all problems. Dr. Rhen expressed a wish that summed up the evening well. She hoped that a farmer at the trade show would report that he supplies his pigs to customers equipped with technology, and that his animals are therefore faring better as a result. Then she would know that progress is being made.