Free "AI chicken soup" has been on the menu this week at Shenzhen University in Guangdong province to celebrate a breakthrough by students in applying artificial intelligence to transform Silkie chicken farming, which has so far increased production by 30 percent.

Silkie chickens are named for their unusually fluffy feathers, which have the texture of silk or satin. They also have black skin and bones, blue earlobes and five toes on each foot — one more than ordinary chickens.

To share the joy of its success at the chicken farm in Chishui, Guizhou province, the Tencent Cloud Artificial Intelligence Class — a program set up by the university and tech giant Tencent — offered the free soup to all teachers and students at the university's cafeteria.

"As Spring Festival is on the way, the farm in Chishui sent hundreds of chickens and eggs to the university, so we decided to have the cafeteria turn these black-boned birds into 'AI chicken soup' before the holiday to give teachers and students a taste of technology," said Chen Sirui, a student in the Tencent class.

Over a span of eight months, about a dozen students from the class have dedicated themselves to developing an intelligent breeding system for the Silkie chickens in Chishui.

The project aimed to address the challenges of large-scale farming, particularly difficulties in monitoring the health and movement of the free-range chickens, which have a breeding history of 300 years.

Recognizing the need for a technological solution to the challenges faced by local farmers, students drew from their previous experience in developing an AI goose breed in Shantou, Guangdong, in 2022.

They successfully enhanced the overall survival rate of the Lion-Head goose (African goose) by 30 percent, utilizing technologies such as "goose body recognition". Building on this success, the students ventured into the Silkie chicken farm in Chishui.

Employing deep learning and computer vision technologies, including target detection, tracking and behavior recognition, they successfully identified and tracked 250,000 Silkie chickens to create a comprehensive database. The result was the world's first intelligent Silkie chicken breeding system.

Operating for more than six months, the system increased the farm's output by 30 percent, yielding an additional 60,000 chickens.

"This achievement not only embraces the centuries-old tradition of free-range farming, but also integrates modern technology. It contributed to the production of high-quality Silkie chickens through the use of cloud technology," said Feng Yuhong, the Tencent class adviser.

Class member Chen highlighted the real-time monitoring capabilities of the intelligent breeding system, which sends valuable data to the chicken farmers' mobile phones.

"The system, although it is still being optimized, aims to further support local farmers in enhancing production and their incomes," Chen said.

Zhang Xiaohong, the university's vice-president, said: "The project not only marks a significant advancement in the application of AI in agriculture, but also underscores the collaboration between academia and industry. This enables students to address real industry challenges, deepen their understanding of technology and develop their practical skills."