A giraffe stirs in the darkness of its enclosure at Marwell Zoo in Hampshire, and tonight, an artificial eye is watching—one that understands what the human keepers cannot see. Researchers from the University of Surrey's Centre for Vision, Speech and Signal Processing are partnering with Marwell Wildlife to deploy an AI-powered nocturnal monitoring system that transforms hidden nighttime footage into actionable insights about animal health and behavior. Over the next three years, the system will learn to spot the subtle signs of illness or distress that might otherwise go unnoticed in the dark hours, helping zookeepers intervene before minor issues become serious problems.
The initiative addresses a genuine gap in animal care. While daytime observations allow staff to monitor visible behaviors, the nighttime world of zoo animals—their movements, rest patterns, and stress responses—has largely remained opaque. The new platform interprets video footage and uses machine learning to flag unusual behavioral patterns, drawing on AI research techniques originally developed to understand human sleep. "We are delighted to be working with the staff at Marwell on this exciting AI project that will deliver the first AI health and welfare monitoring system focused on zoo animals," said Professor Kevin Wells, project lead at Surrey.
Giraffes and red river hogs will be the first species to benefit from the system as researchers develop and refine it across different animal types. The technology promises benefits that extend beyond immediate health alerts. By tracking how animals interact with their environments—how lighting, heating, or feeding times influence behavior—zoos gain a deeper understanding of what genuine animal thriving looks like. This data-driven approach strengthens the evidence base for modern zoo management, transforming intuition into quantifiable knowledge.
Marwell Wildlife has long positioned itself at the forefront of animal welfare innovation, having pioneered evidence-based assessment frameworks and naturally positive life experiences for the animals in its care. Laura Read, the organization's Chief Executive, emphasized this commitment: "In a modern zoo, thriving animals and habitats are absolutely paramount." The partnership reflects a broader philosophy that technology and conservation expertise can combine to push welfare standards higher than they might otherwise reach. A dedicated KTP Associate—a graduate researcher—will work full-time on the project, gaining hands-on experience in applying AI and data science in real conservation settings, ensuring the next generation of researchers understands both the technical and ethical dimensions of this work.
What makes this project distinctive is its potential to scale beyond a single zoo. Researchers and staff envision eventually monitoring other species and deploying the system across wildlife parks, conservation sites, and even livestock operations. The technology could become a new standard in how we care for animals under human supervision, turning the night from a period of mystery into one of informed stewardship. In an era when data science often feels distant from lived experience, this partnership grounds AI in a fundamentally compassionate goal: ensuring that the animals in our care are truly thriving.