“Alexa, how are my chickens?” Ok, it doesn’t quite work like that, but Cargill is researching how audio data collected from birds on poultry farms can be used to help improve animal health and welfare.

Using machine learning, a form of artificial intelligence (AI) technology, Cargill data scientists have been analysing chicken vocalisations to try to identify patterns and changes that may signal a health or welfare problem before it’s apparent to the human eye.

“Early detection of health concerns will allow for intervention to diagnose the issues and take corrective actions,” said Dries Tromp, digital lead for Cargill’s poultry business. “We are developing an approach that is focused on ensuring chickens are healthy and receive great care, which is essential to the long-term sustainability of the business.”

Cargill is conducting a pilot at two poultry farms in China, where the company operates a fully integrated poultry business. In Canada, Cargill is also using video to record data on a poultry farm that supplies its protein business. There, the company is looking at how the birds’ movements may indicate health changes over time.

In both cases, the teams have had to troubleshoot issues that come from collecting data in the farm environment. For example, they realised that natural temperature fluctuations, dust and other conditions can impact the hardware needed to collect data.

“Developing systems that work is a lengthy progress of trial and error,” said Tromp. “We are still learning and testing and are staying agile in our approach.”

The company also brought together farmers, animal science leaders and technology experts to help craft solutions.

“Together we can harness the power of digital technology to deliver safe and transparent poultry products to consumers,” said Tromp.