Artificial intelligence that could improve the welfare of farmed chickens by eavesdropping on their screeching could become available within five years, researchers say.
The technology, which detects and quantifies emergency calls made by chickens housed in massive indoor sheds, correctly differentiated emergency calls from other barn sounds with 97% accuracy, new research suggests. A similar approach could eventually be used to raise welfare standards in other farm animals.
Every year around 25 billion chickens are kept around the world – many of them in huge barns, each housing thousands of birds. One way to judge the well-being of such creatures is to listen to the sounds they make.
“Chickens are very vocal, but the distress call is usually louder than the others, and is what we would describe as a pure tonal call,” said Alan McElligott, an associate professor of animal behavior and welfare at the City University of Hong Kong. “Even for the untrained ear, it’s not that hard to pick them out.”
In theory, farmers could use the call of chickens to measure distress and enrich their housing where necessary. However, in commercial flocks with thousands or tens of thousands of chickens, it is impractical to use human observers. For starters, their presence could further stress the flock, but with so many birds, it’s impossible to objectively quantify the number of emergency calls, McElligott said.
Instead, his team has developed a deep learning tool to automatically recognize chicken distress calls from recordings of intensively farmed chickens. The tool was trained using recordings already manually classified by human experts to determine what type of sound they represented.
According to an evaluation published in the Journal of the Royal Society Interfacethe algorithm correctly identified 97% of emergency calls.
“Our ultimate goal is not to count emergency calls, but to create conditions in which the chickens can live and be less distressed,” said McElligott, who estimates the technology could be deployed commercially within five years.
Before that happens, the team must make sure the recording equipment works in different types of chicken sheds, and test it in farms with higher or lower welfare standards to confirm that the measurements correlate.
Convincing farmers to adopt the technology may be relatively easy. Previous research by McElligott found that distress calls from young chicks could predict the amount of weight gain and number of deaths in the entire flock over their lifetime.
“Sometimes it’s hard to convince the farmers to produce these animals for a fixed price for supermarkets and everyone else to use technology to improve their welfare,” McElligott said. “But we’ve already shown that emergency calls are a good indicator of mortality and growth rates, and this is one way to automate the process.”
Similar technology could be developed to track other farm animals, especially pigs or turkeys, which are also often housed indoors and are very noisy, he added.
The RSPCA broadly welcomed the investigation. “Technology like this can be incredibly helpful in monitoring and improving farm animal welfare, but we wouldn’t want it to replace physical inspections or reduce contact between livestock farmers and birds as it could lead to a loss of skills in the field.” livestock, or birds that are more difficult to handle,” said a spokesperson.
“In addition, expressing distress is only one indicator of wellbeing, but there are a number of other physical factors such as lameness and leg burns that farmers should also be aware of.”