Understanding how pigs interact socially is a cornerstone of successful animal management. Yet objectively measuring these interactions has historically been challenging.
A groundbreaking study by The Roslin Institute, PIC, and other partners now delivers a fresh, data-driven approach to animal behaviour. It demonstrates that artificial intelligence (AI) tracking, combined with social network analysis (SNA), can reveal hidden social structures among pigs in ways never before possible.
This innovative method opens major opportunities for management practices, health monitoring, and genetic selection strategies in pig farming.
The research at a glance
Researchers implemented a fully automated monitoring system on a PIC breeding farm, housing pigs in six pens with 16 – 19 animals each. Using overhead cameras equipped with AI, they continuously tracked individual pigs via ear tags.
Data collected included:
- Ear-tag identification
- Time and location data (XY coordinates)
- Posture (standing, lying, sitting)
- Feeding and drinking duration
- Proximity to other pigs
Based on this data, researchers constructed detailed social interaction networks—mapping how frequently and closely pigs interacted over time.
Key findings: Social order is not random
One of the study’s most compelling insights was that pig social structure becomes increasingly organized as the animals grow.
While pigs initially exhibit relatively random interactions, over time, a clear hierarchy emerges. Certain pigs consistently become “central nodes”—interacting broadly and playing pivotal roles in the group’s stability.
While pigs initially exhibit relatively random interactions, over time, a clear hierarchy emerges. Certain pigs consistently become “central nodes”—interacting broadly and playing pivotal roles in the group’s stability.

In practice, this means:
- Social hierarchy forms naturally: Some pigs dominate social space early.
- Clustering increases over time: Pigs create tighter, more predictable subgroups.
- Behavioural roles are stable: Individual pigs tend to maintain their social positions, opening new avenues for trait selection.
Practical implications for pig farming
This research offers several promising pathways to improve welfare, productivity, and herd management:
1. Early identification of social outliers
AI and SNA can help producers detect pigs that are unusually aggressive, overly passive, or socially isolated traits often linked to health or welfare issues. These insights allow for proactive interventions, such as tailored enrichment or re-grouping.
2. Enhanced group composition
Understanding how pigs form structured networks enables more strategic pen grouping, reducing stress-related behaviours like fighting and tail biting.
3. Optimized pen design
Data-driven insights into how pigs occupy space can inform better pen layouts that support natural group dynamics and minimize conflict.
4. Digital Phenotyping for breeding
Perhaps most excitingly, stable social behaviour traits could become part of genetic selection programs. Imagine selecting pigs not only for growth and feed efficiency but also for social compatibility traits that promote welfare and performance.
Practical Example: Some pigs naturally move between different groups and help maintain calm and connection within the pen. These “keystone pigs” play an important role in keeping the group stable. Identifying and selecting animals with these traits could help build stronger, more resilient herds.
Bridging technology and animal behaviour
This research highlights the power of combining machine learning, behavioural analysis, and genetics.
While still in early stages of commercial application, the approach offers a forward-looking model for precision livestock farming.
Uniquely, this system enables scalable, real-time monitoring of social behaviour — without human bias or manual observation. It opens the way for AI solutions that not only monitor pigs but also interpret their behaviour to optimize care, grouping, and management strategies.

PIC’s role in pioneering practical solutions
As a co-author of the study, PIC continues to lead the industry in translating science-backed insights into farm-level solutions.
From advanced genetics to innovative monitoring technologies, our mission remains clear: helping pork producers maximize animal well-being, performance, and profitability.
Final thoughts
The intersection of AI and animal science is no longer theoretical—it is already reshaping modern pig farming.
By leveraging technologies like social network analysis, pig producers and advisors can unlock a deeper understanding of animal behaviour—leading to more sustainable, welfare-focused, and efficient operations.
“Pigs aren’t just growing—they’re communicating. This technology lets us listen.”