EPIC is the Scottish Government-funded Centre of Expertise in Animal Disease Outbreaks, now entering its 19th year in operation, and BioSS colleagues working in Animal Health and Welfare have been central to EPIC’s work from the beginning.  EPIC’s remit is to support the Scottish Government in its work to maintain the health status of livestock in Scotland, in particular addressing the risks associated with exotic notifiable diseases. Each year, EPIC organises a conference to support interactions between researchers, science/policy brokers and policymakers, attracting scientists from all over the UK and beyond, and policy staff from Scottish Government, Defra, and other devolved administrations, plus operational colleagues from the Animal and Plant Health Agency.  BioSS colleagues had a prominent role in the 2025 meeting, held over two days at the beginning of March in Edinburgh.  The theme of the conference was: "Building a Resilient Future for Animal Health and Welfare: Data-Driven Innovation in Policy and Practice", with a strong, but not exclusive focus on the potential use of Artificial Intelligence (AI) to support provision of scientific advice to policy-makers.  AI is likely to provide new opportunities to process data and information, while innovations in precision agriculture for livestock have potential to generate much richer datasets in both the public and private spheres: we thought it useful to think of these complementary developments together.

BioSS staff presented their work to develop mathematical and statistical methodologies to make best use of what demographic and epidemiological data are available, working around limitations in the availability, quality and provenance of relevant datasets. Dr Giles Innocent outlined an approach using synthetic data sets. By not using the original data, many issues arising from data protection regulations, such as the need to limit access, are avoided. In addition, such approaches allow multiple data sets to be generated, where these have similar characteristics to the true data, permitting the robustness of model outputs to be readily tested. Dr Stephen Catterall presented some of his work on modelling the spatial spread of disease when the spatial distribution of the disease host is unknown. This, not uncommon, scenario does not align well with the requirements of conventional modelling approaches; he presented alternative methods that make use of auxiliary datasets to work around the lack of host distribution data.

Image

Giles Innocent presenting at the EPIC 2025 Conference

The conference welcomed presentations from Professor Christos Tachtatzis and Dr Jess Enright, who are members of the current Specialist Advisory Group on AI convened by our funder, RESAS, under the chairmanship of BioSS Director Mark Brewer, to advise on the ways that AI might support the scientific priorities of RESAS in environmental and agricultural policy. Talks describing current cutting edge uses of AI on livestock premises, and the wide potential uses of AI were balanced with talks from Tom Wilkinson, the Chief Data Officer for Scottish Government and Miruna Clinciu, from the Edinburgh Centre of Robotics, stressing the unreliability of current AI technologies, the importance of an ethical approach to AI development and use, and the central importance of the ‘Right to Explanation’ and methodologies to promote AI explainability.  Professor Iain McKendrick, from BioSS, who is the Director for Quantitative Innovation in EPIC, contributed to a panel discussion about the future opportunities and barriers to uptake in use of data-intensive methods to meet the needs of the livestock sector and associated policy audiences. 

Iain expressed the belief that AI methods will rapidly become valuable tools, delivering ‘automated cognition’ to process data pipelines where the properties of the data and the objectives of the exercise are well defined.  In such situations, AI will support the collection and collation of data sets which, although in principle available already, are too expensive in terms of person-time to collect and process.  As AI methods develop, methodologies which help humans understand the basis of decisions that are being made on our behalf will become ever more important.  Iain suggested that these are areas with which EPIC is likely to have to engage in the near future.  BioSS has well developed plans to engage with AI as a research topic, and we have identified that, in both these respects, a common denominator is a necessity to incorporate aspects of statistical thinking into methodological development, expertise that BioSS will be well positioned to deliver.

Image

BioSS presenting at EPIC 2025: Iain (Top-Left); Stephen (Bottom-Left); Giles (Right)

Photo credits: Lewis James Houghton