Computer vision based assessment of animal pain

Rapid progress in computer-based automation approaches created new opportunities - and posed new challenges- for research and applications in animal wellbeing. Innovative systems are being developed to support pet owners, animal caregivers, and veterinarians, particularly in monitoring of animals across diverse environments. 

However, these developments also raise concerns akin to those in automatic human behavior analysis, where automation risks diminishing the human touch and introduces ethical and practical questions about the responsible use of such systems, both in the short and long term. 

 

In this talk, I draw parallels between computer analysis of human and non-human animal behavior, providing examples from our interdisciplinary work in equine and canine pain estimation, as well as ongoing research at the newly established AI and Animal Wellbeing Laboratory at Utrecht University, where computer scientists collaborate with veterinary experts and ethologists. Based on these examples, I raise a number of technical and societal challenges in this research area, and argue for the inclusion of animal rights within the scope of AI ethics.