Artificial intelligence may eventually help us to understand when animals are in distress or showing other emotions, at least according to researchers profiled in Science.
The Intellipig system, for example, is being developed by scientists from the University of the West of England Bristol, and Scotland’s Rural College. It examines photos of the faces of pigs and notifies farmers of any signs of pain, illness, or emotional distress.
A team at the University of Haifa, one of the teams behind facial recognition software used to find lost dogs, is now training AI in order to identify signs of discomfort. Dogs share 38% of their facial movements with humans.
These system rely on humans to do the initial work in identifying the meanings for different animal behaviors. (Usually based on long observations of animals in different situations). A researcher at the University of Sao Paulo recently experimented using photos of horses faces before and following surgery, and before and following painkillers. He trained an AI system to focus only on their eyes, mouths, and ears. The system was able, with an 88% rate of success, to learn what signs might indicate discomfort.
