KI entschlüsselt Tierkommunikation: Forscher verstehen die Sprache von Schweinen | NDR Info

Summary

Researchers are using artificial intelligence to assess pigs’ emotional states from their grunts. An international team recorded more than 1,000 hours of pig vocalizations on different farms and developed an algorithm that distinguishes positive, happy states from negative, anxious ones with 92% accuracy.

The aim is to identify calls from individual pigs and document how their welfare and stress levels vary across situations, including being alone or with other pigs, having more or less space, and being transported to slaughter.

Researchers say pigs can express as many as 19 different emotions. Measurements also found negative emotions in as many as 80% of cases.

Eiken Struve’s conventionally managed farm performed better than expected despite limited space per pig: researchers recorded many positive calls and few negative ones. They suggest that regular, attentive care from people may help explain the result.

Overall, the findings indicate that pigs’ welfare can be assessed through their vocalizations, and that attentive human care may measurably improve it.

Researchers and Sources Featured

• Eiken Struve — Farmer whose pigs were studied.
• An unnamed international team of scientists — Recorded pig vocalizations and developed the AI algorithm.