Researchers are now reportedly experimenting with artificial intelligence (AI) to decipher the facial expressions of animals to understand their emotions. As per the report, several research groups are now using AI models with computer vision to analyse emotions and signs of pain and distress in farm animals. This is one of the niche use cases of the technology, but provided enough high-quality data, a large language model can figure out in real time whether an animal is in need of assistance.
Researchers Using AI to Decipher Animal Emotions
According to a Science.org report, several research groups are studying whether AI can be reliably used to monitor the well-being of farm animals. One such example is the Intellipig system, developed by researchers at the University of the West of England Bristol (UWE) and Scotland’s Rural College (SRUC). The project, which is detailed on a web page of the university’s website, is currently in its beta testing phase.
Researchers have now reportedly implemented the Intellipig system in farms and are using AI to monitor hundreds of pigs. The report claims that every morning, a picture of every pig is taken before they are given food, and based on the AI-based identification, each pig is given specific food.
Additionally, the AI also runs a separate inference to analyse the facial data and understand if there are any signs of pain or distress. In case there are, the AI alerts the farmer, who can then take a closer look and determine if the situation requires any intervention. Notably, the researchers have also claimed that the system can eventually analyse the emotions of these animals as well.
Notably, deciphering health issues in animals from facial expressions and other behavioural patterns is not a new science. A lot of literature exists on this, and humans trained in the subject matter have been using the knowledge to understand when an animal might be suffering from an ailment or is emotionally not well.
However, as per the report, relying on humans is not a sustainable method, as they need an average of 100 seconds to identify the various facial muscles. On the other hand, AI can do the same in near real-time at a large scale. Researchers are not planning to establish an autonomous system with AI in charge, as the technology is currently in a nascent stage.
