Mitigating artificial intelligence-related animal welfare risks in animal farming

In animal farming, artificial intelligence (AI) is increasingly promoted as a means to enhance productivity, reduce greenhouse gas emissions, and improve animal welfare. However, discussion of the animal welfare risks arising from these technologies, particularly emerging generative AI and computer vision tools, remains limited. To address this gap, we conducted a scoping review to map technology-related welfare risks in animal farming. Firstly, by using 72 retained papers, we created a key word map, indicating that while precision livestock farming and animal welfare were commonly linked, AI tools were rarely discussed alongside ethical considerations. Secondly, using a subset of 52 papers that specifically discussed technology related welfare risks or mitigation strategies in detail, we examined how AI may drive or amplify each risk. Across the four broad risk categories identified, we selected five individual risks for which mitigation strategies are particularly lacking: (1) embedded value hierarchies; (2) erosion of traditional farmer skills; (3) shift of farmer identities; (4) objectification, instrumentalization, and speciesism; and (5) humane washing. Thirdly, drawing in part on insights from other disciplines, we propose potential solutions to these five risks. We conclude that coordinated action among technology developers, farmers, and policymakers is necessary to ensure that responsibility for animal welfare is matched by the technical and economic capacity to act.

University of Southampton
University of Exeter
Johns Hopkins University
Central European University
University of Cambridge
New Jersey Institute of Technology
University of British Columbia