Recent outbreaks of mpox, avian influenza, dengue, and Ebola have highlighted the need for faster, more integrated approaches to disease detection and response. Public health agencies have access to vast amounts of data, but much of it is fragmented and difficult to analyse quickly.
‘Public health authorities already have access to enormous amounts of information, but the challenge is turning that information into reliable intelligence quickly enough to make a difference,’ said Professor Richard Maude, Principal Investigator of PANDAI, Professor of Tropical Medicine at the Nuffield Department of Medicine and Head of Epidemiology at the Mahidol Oxford Tropical Medicine Research Unit (MORU).
‘PANDAI will use AI to bring together diverse sources of data, assess what emerging threats mean for different places and populations, predict how they may develop, and provide the evidence needed to make faster and better-informed public health decisions.’
The World Health Organization (WHO) Health Security Division/Europe will define what the system needs to do and how it should be structured, and will ensure that the PANDAI platform integrates smoothly with existing frameworks, such as the Epidemic Intelligence from Open Sources (EIOS) initiative. The WHO will also ensure the system can be used by national ministries of health and global policymakers.
Dr Marc-Alain Widdowson, World Health Organization European Region said ‘This initiative will make a real difference to how the public health community can act on data, by bringing together far more sources and perspectives than we've had access to before. By building on existing WHO systems and expanding use of AI, we'll be able to pick out and prioritise the most important signals faster, build a fuller picture of the risk, and quickly recommend the right control measures for the situation at hand serving member states.’
At the core of the Observatory is the creation of a cutting-edge AI model, one of the first epidemiology-informed domain-specific foundation models (EDFMs) for infectious disease surveillance and epidemic intelligence. It will integrate disease surveillance, environmental, health system, veterinary, preparedness capacity, mobility and pathogen genomic data to provide experts/public health authorities with AI-driven insights.
Professor Melinda Mills, co-investigator from Oxford Population Health said ‘PANDAI brings together data on population behaviour, mobility, health systems and demographics to better understand how diseases spread and how people respond. By focusing on strong communication, integrating behavioural social science, attention to diversity and stakeholder perspectives into AI models from the outset, we can build tools that are not only more accurate, but also more equitable, trusted and useful for public health decision-making.’
The PANDAI platform will undergo rigorous, multi-country validation across three distinct transmission pathways: vector-borne arboviruses (such as dengue and West Nile virus), respiratory viruses (exemplified by seasonal and novel avian/swine influenza strains like H5N1), and direct-contact infections (mpox). This will help ensure that the platform's framework can be generalised to future emerging infectious disease threats, including the unknown "Disease X”.
Public health officials will be able to conduct secure "what-if" simulations, assessing the real-world impact of interventions, such as localised lockdowns or prioritised vaccination schedules, prior to field deployment.
The project will also address the ethical and legal challenges of using AI for public health surveillance, with measures to protect privacy, minimise the use of personal data and reduce the risk of bias. A compliance work package, led by the Luxembourg National Data Services, will institute protocols and assessments to mitigate algorithmic bias while ensuring compliance with the EU AI Act, proving that pioneering AI innovation can seamlessly coexist with robust data privacy protections.
PANDAI consortium partners
The PANDAI project brings together nine partner organisations spanning public health, artificial intelligence, data science, technology, governance, and civil society across seven countries:
|
• University of Oxford (United Kingdom) – Coordinator |
• Technovative Solutions Ltd (United Kingdom) |
©
John Cairns