Israeli AI Detects Pain in Horses by Reading Their Faces

Horses are masters at hiding their pain, making it nearly impossible for even experienced veterinarians to know when they're suffering. But a breakthrough AI system from Israel can now read their faces to detect discomfort before it becomes a serious problem.

Researchers at Tel-Hai University developed SHIC-XE, an artificial intelligence tool that analyzes video footage to identify signs of pain in horses. The system focuses on subtle cues like ear position and cheek muscle tension that humans often miss.

Led by Marcelo Feigelstein, head of the Artificial Intelligence Systems Engineering program, the international team solved a tricky challenge. Traditional AI systems that analyze images struggle with video because the subject keeps moving. Their two-dimensional attention maps become jumbled and hard to interpret when an animal shifts position or the camera angle changes.

The breakthrough came when researchers projected the AI's observations onto a fixed three-dimensional model of a horse's face. Now the system maintains consistent anatomical tracking no matter how much the horse moves or where the camera is positioned. The explanations match up with expert veterinarian assessments.

This isn't Feigelstein's first rodeo in animal welfare AI. His team has already created similar pain detection tools for cats, dogs, rabbits, sheep, and cattle. Animals across species have evolved to hide suffering as a survival instinct, but that makes modern care incredibly difficult.

The research team included scientists from universities in Bern, Milan, São Paulo, and Newcastle, alongside Israeli institutions. Their findings appeared in the International Journal of Computer Vision.

Why This Inspires

Right now, SHIC-XE simply identifies whether a horse is in pain. But the team is already training the next version to assess pain intensity and potentially pinpoint its source. That could revolutionize how we care for animals who can't tell us what hurts.

The technology could work anywhere horses live and work. Imagine farms where monitoring systems catch illness early, veterinary hospitals with real-time pain assessment, or transport vehicles that alert drivers when animals become distressed. The researchers still need to create practical applications, but the foundation is solid.

Every year, countless animals suffer in silence because we simply can't read their signals. This technology gives them a voice.