Can AI read pain in your cat's face? (tailstory-app.com)

Veterinary researchers have developed a method for detecting pain in cats using the Feline Grimace Scale, which scores facial expressions across five specific features: ears, eyes, muzzle, whiskers, and head position. In a recent study, an AI model trained on 3,447 annotated cat photos achieved an impressive 95.5% accuracy in assessing pain from these images. This advancement is significant because it provides a reliable, automated tool for veterinarians and pet owners to identify pain in cats, a species notoriously adept at masking discomfort. However, a 2025 study tested various general chatbots, including ChatGPT and Gemini, to see if they could accurately score the Feline Grimace Scale based on uploaded cat photos. Unfortunately, all chatbots consistently undervalued the cats' pain levels compared to a veterinary expert, underscoring the importance of specialized models for specific tasks. This distinction highlights the limitations of using general AI for sensitive evaluations and reinforces the need for purpose-built tools in veterinary care, ensuring that cat pain is recognized and addressed effectively.