AI-enabled animal welfare monitoring put to the test
Research has shown that artificial intelligence-enabled monitoring systems can enhance animal welfare oversight at a multi-species abattoir when integrated within structured governance, training and data management systems to achieve sustained improvements.
A project funded by AMPC, and conducted with partners Impetus Animal Welfare, deployed an AI-enabled monitoring system with CCTV and sensor devices at a commercial multi-species abattoir. The system was installed, trained and integrated across calves, sheep and goats.
Measurable welfare indicators were captured through a combination of computer vision, sensor devices, manual event logging and a structured review process.
The project demonstrated that AI-enabled monitoring systems are capable of providing continuous visibility of defined animal welfare indicators, supporting compliance activities and strengthening the traceability of corrective actions.
AMPC Project Officer for Markets and Integrity Dr Sarah Babington said animal welfare monitoring in Australian red meat processing facilities continued to be under increasing public scrutiny which was influencing customer-drive assurance requirements as well as animal welfare regulations.
"Collectively, that is reshaping expectations around transparency, documentation and operational control at processing facilities," Dr Babington said.
"In particular, the staged introduction of mandatory video surveillance under the Australian Animal Welfare Certification System (AAWCS) has accelerated the interest in, and adoption of, surveillance technologies across the sector.
"While installation of video surveillance systems like CCTV are common in processing facilities, the effective integration of these types of technologies into operational decision making remains highly variable."
Animal welfare monitoring at processing facilities typically relies on a combination of manual monitoring activities, paper-based record keeping and periodic customer and regulatory audits.
"These systems, while often compliant, are inherently episodic and impacted by subjective interpretation, individual capability and documentation quality," Dr Babington said.
"The introduction of digital monitoring systems, particularly those incorporating artificial intelligence, computer vision and sensor-based devices offers the potential for more continuous, objective, and structured data collection. However, technology alone does not guarantee improved welfare outcomes. Its value depends on the organisational systems within which it operates, including leadership alignment, training, corrective action processes, and data governance."
The results confirm that AI computer vision systems alone are not sufficient to capture the full complexity of welfare performance at processing facilities; however, when integrated appropriately, they materially enhance operational transparency and management capability, Dr Babington said.
Impetus Animal Welfare chief executive officer Dr Michael Patching said livestock areas at processing plants had zones that contained risks that could only be managed through continuous monitoring.
"Technology now allows us to quantify what is happening all day long," he said.
"We super-power our human review when we link data from our electric prodders, stunning equipment and computer vision to camera footage.
"This provides defensibility of our businesses and industry, and proactive improvement and smarter investment."
