# A Comparison of Artificial Intelligence and Human Observation in the Assessment of Cattle Handling and Slaughter

## TL;DR

This study compares AI and human assessments of cattle handling in slaughter plants and finds they are highly similar for most outcomes.

## Contribution

The paper introduces AI as a reliable tool for evaluating cattle handling outcomes in slaughter facilities.

## Key findings

- AI and humans showed high similarity in identifying stunning, electric prod usage, and falls.
- AI was effective at identifying questionable handling events for further review.
- Human-AI collaboration could improve detection of suspicious animal handling events.

## Abstract

Animal welfare is a critical component of food animal production. It is important to ensure that animals are provided a good quality of life, up to and including their death. In addition to following governmental regulations for humane handling, companies that slaughter animals also perform internal assessments and often employ the use of technology (e.g., remote video auditing) to enhance their auditing programs. Artificial intelligence (AI) was introduced to enhance the automation of the process and provide more robust oversight (i.e., 24/7 vs. ad hoc). The use of AI to evaluate cattle handling outcomes was compared to the use of human observation to provide insights regarding the accuracy of using AI systems. The results of this project demonstrated that AI could effectively identify cattle handling outcomes such as falling, stunning, and electric prod usage. Humans were also able to accurately identify these outcomes. In some instances, particularly related to questionable animal handling events (i.e., unusual or suspicious activity that could be considered egregious), it would be valuable to use a human-enhanced AI approach, as was conducted in the plant that participated in this study. Artificial intelligence technology could be implemented in slaughter plants to enhance animal welfare evaluation.

Slaughter facilities use a variety of tools to evaluate animal handling, including but not limited to live audits, the use of remote video auditing, and some AI technologies. The objective of this study was to determine the similarity between AI and human evaluator assessments of critical cattle handling outcomes in a slaughter plant. One hundred twelve video clips of cattle handling and stunning from a slaughter plant in the United Kingdom were collected. The AI identified the presence or absence of: Stunning, Electric Prod Usage, Falling, Pen Crowding, and Questionable Handling Events. Three human evaluators scored the videos for these outcomes. Four different datasets were generated, and Jaccard similarity indices were generated. There was high similarity (JI > 0.90) for Stunning, Electric Prod Usage, and Falls between the evaluators and the AI. There was high consistency (JI > 0.80) for Pen Crowding. There were differences (JI ≥ 0.50) between the humans and the AI when identifying Questionable Animal Handling Events but the AI was adept at identifying events for further review. The implementation of AI to assist with cattle handling in a slaughter facility environment could be an added tool to enhance animal welfare programs.