Artificial intelligence supports pig welfare

A combination of tracking technology and social mapping that reveals how pigs build relationships over time could inform welfare strategies.

Artificial intelligence (AI) has been combined with techniques to map animals’ interactions, to gain insights into pigs’ social relationships.

Scientists at the Roslin Institute, in collaboration with industrial partners PIC and an international research team, examined how pigs associated with one another over time by using technology to analyse their proximity during social interactions.

Findings reveal that as pigs spend more time together, their social interactions become more structured. Some pigs emerge as central figures within their groups, indicating the development of a social hierarchy. 

These insights can help breeders and farmers better understand animal behaviour and improve breeding and management practices.

A deeper understanding of pigs’ social interactions can also aid in reducing stress-related behaviours, such as tail biting, a common health and welfare issue that can lead to major industry losses.

Beyond monitoring individual health or activity, the research demonstrates how data-driven approaches can support more informed decisions on breeding, housing and welfare strategies, the research team explains.

Image of pigs numbered, as seen by the AI camera used in the study. Lines show each pig's sitting or standing position.
The 2D camera images provided by the automated recording systems, which show the ear tag, position and posture, e.g., standing (with red line) and sitting (blue line), for each pig within the pen.

Mixed methods

The study, conducted on a commercial pig breeding farm in the United States, demonstrates how emerging technologies can be applied to modernise animal husbandry. 

An automated monitoring system tracked pigs housed in indoor pens, collecting data on individual animals using cameras and movement sensors to record the pigs’ posture and location in real time.

The use of artificial intelligence eliminated the need for manual observation, which is typically time-consuming and labour-intensive.

Social analysis

In order to map each pig’s social connections, researchers then applied a method traditionally used to study individual relationships in human social groups, known as social network analysis.  

Key indications of social connection, such as how central a pig was in the network or how connected it was to others, increased over time, suggesting that pigs gradually form more stable and structured social relationships as they grow. 

This shift highlights the dynamic nature of pig social structures, which were previously difficult to monitor at such detail and scale.

This research was published in Animals, in collaboration with colleagues from Scotland’s Rural College and Iowa State University. 

As the farming industry continues to adopt smart technologies, the ability to analyse animal behaviour at both the individual and group level represents a significant step forward. 

"This study not only highlights the power of AI in agriculture but also emphasises the importance of considering social behaviour in animal management.

Professor Andrea Doeschl-Wilson, Group Leader in infectious disease genetics and modelling

The Roslin Institute receives strategic investment funding from the Biotechnology and Biological Sciences Research Council and it is part of the University of Edinburgh’s Royal (Dick) School of Veterinary Studies.

This article was published on