Delivering on the promise of automated laboratory animal welfare monitoring: Data,technology and validity

Monitoring laboratory animal welfare in an effective and timely manner is crucial if we are to meet the promise we give when we sell the 3Rs principle as a guarantee that animal suffering is minimized in biomedical research. As the social licence for doing experiments on animals relies on our meeting this ambition, this is no small claim. It is also no small task, as it requires that we understand what signs to look for in order to measure animal welfare, and that we have suitable means to collect and process information at critical timepoints. In other words, we need knowledge and we need technology.

Writing this commentary in early 2026 is an opportunity to reflect on how far we (have not yet) come in generating the knowledge that is needed to finetune technology to serve laboratory animal welfare. As an animal welfare scientist with a farm animal background, it has long intrigued me how far behind we are in the laboratory animal community when it comes to both generating this knowledge and developing the technology. Ten years ago, a commissioned review on precision livestock farming for dairy cows could report on technologies being tested in as early as the 1980s. ¹ In 2026, dairy farmers monitor cow health and reproductive status using automated measures of activity, food consumption and temperature in their animals. In contrast, we still consider the clinical scoresheet and the (not necessarily trained) eye of the research student or, at best, research technician standard practice to monitor laboratory animal health and welfare.

But writing this commentary at this moment is also an opportunity to consider what is, hopefully, no longer far from becoming reality in laboratory animal rodent welfare. The submission of a Perspective paper on laboratory animal welfare assessment methods prompted a discussion with the Editorial Board of where the field is and how to best bring together state-of-the-art knowledge for the benefit of the community, including those who are responsible for ensuring laboratory animal welfare in research facilities. The decision was to commission two additional papers, resulting in the present publication of three Perspective papers on the topic.

Talbot et al. ² describe how multi‑parameter, AI‑enabled monitoring can potentially capture complex behavioural and physiological changes earlier and more objectively than traditional single metrics. They build on the large interdisciplinary multisite project Severity assessment in animal-based research, generating data on model severity through the systematic monitoring of different procedures and disease models. In their paper, they illustrate how such diverse data can be combined into composite severity scores, for an evidence-based and model-specific approach.

Drawing on the combined expertise from the European-wide project Improving biomedical research by automated behaviour monitoring in the animal home-cage, Tremoleda et al. ³ explore the potential of home-cage monitoring systems. While these allow continuous, non-invasive monitoring, their successful use depends on overcoming challenges in standardization, interoperability and data management – but also in understanding what the data actually mean.

Building on her long experience of devising, implementing and critically assessing animal welfare research approaches, Mason ⁴ provides the conceptual anchor, reminding us that any technological advance is only as meaningful as the construct validity of the welfare indicators it employs. Her analysis highlights a central tension: while digital tools promise precision and scalability, they can inadvertently amplify flawed assumptions if not grounded in validated measures of animals’ affective states.

Together, these perspectives showcase the potential of digital and AI tools in collecting and processing data on laboratory animal welfare and highlight that only when paired with conceptual rigour and thoughtful implementation can they truly enhance animal welfare and scientific quality.