Saves clucking. AI can be used to alleviate the suffering of chickens on intensive farms

The scientists believe it will be easy to convince the researchers as the diagnosis could help produce heavier and healthier animals.

Over the past few years, with increasing awareness of animal careespecially those often used as human food and included in intensive farming systems, many of these practices have been interrogated and questioned. Now science seems to have arrived at a way to use artificial intelligenceprovide more security for the welfare of farmed chickens through hear your cackle is a technology that could be available within five years, the researchers note.

Function that detects and quantifies requests for help made by chickens housed in huge sheds, correctly distinguishes cries for help from other sounds using 97% accuracy, suggest investigation. A similar approach can be used to raise welfare standards for other farm animals, points out The keeper.

Annually about 25 billion chickens are bred all over the world – many of them in huge pavilions with poor conditions, each of which is home to thousands of birds. One way to assess the welfare of such animals is to hear sounds what are they doing.

“Chickens are very loud, but the cry for help is generally louder than others, and is what we would describe as a pure tonal cry,” he said. Alan McElligott, Associate Professor, Department of Animal Behavior and Welfare, City University of Hong Kong. “Even to the untrained ear, it’s not very difficult distinguish them.”

Theoretically, farmers could use the clucking of chickens to evaluate their distress level e improve your conditions when needed. However, on commercial farms with thousands or tens of thousands of chicks, such human analysis is not practical. On the one hand, his presence can further upset the animals.

Instead, McElligott’s team developed a tool to deep learning is able to automatically identify the distress calls of chickens from the records of chickens raised in intensive care. The tool was trained using records are already classified manually by human experts to determine in advance what sound they represent.

According to an estimate published in Journal of the Royal Society Interface, the algorithm correctly identified 97% of the distress calls. “Our ultimate goal is not to count calls for help, but to create conditions in which chickens can live and have reduced amount distress signals,” explained the researcher, who believes the technology could be commercially deployed within five years.

Before this happens, the team must ensure that the recording equipment is working in various types of premisesas well as testing on farms with more or less high welfare standards to confirm correlated readings.

Convincing breeders to adopt this technology may be relatively easy, as McElligott’s previous research has shown that distress calls from young chicks can predict amount of weight gained this is number of deaths throughout the group during its lifetime.

“Sometimes it can be difficult to convince breeders who work with these animals at a fixed price for supermarkets and everyone else to adopt the technology to improve their welfare,” McElligott said. “But we have already demonstrated that requests for help good indicator mortality and growth rates, and this is the form automate the process“.

A similar technology could be developed to monitor other animals He added that livestock, especially pigs or turkeys, who are also often kept indoors, are very noisy.