WOAH Launches Eight AI and Data Science Challenges to Strengthen Global Animal Health Surveillance

The World Organisation for Animal Health’s new datathon initiative targets earlier disease detection, outbreak forecasting, antimicrobial-use monitoring, trade-risk analysis and AI-powered veterinary product intelligence. The challenges could help shape the next generation of digital tools for animal-health authorities, pharmaceutical companies, diagnostic developers and livestock producers.

GLOBAL | 13 October 2026 — The World Organisation for Animal Health (WOAH) has published eight analytical challenges under its WOAH Datathon initiative, outlining how data science, artificial intelligence (AI) and improved data integration could strengthen animal-health surveillance and support decision-making across countries and species.

The challenges address some of the persistent weaknesses in global animal-health intelligence: delayed outbreak notifications, fragmented information, differences in national surveillance capacity, limited integration of trade and disease data, antimicrobial-use monitoring gaps and the difficulty of extracting actionable information from large volumes of unstructured reports.

The programme also identifies opportunities to build structured databases of veterinary antibiotic products and to improve the analysis of long-term disease patterns. Its scope spans companion and production animals, with potential applications in poultry, swine, cattle and other livestock sectors, depending on the datasets and diseases examined.

For the animal-health industry, the initiative is relevant beyond public-sector surveillance. Better-integrated information could support disease-risk assessment, diagnostic planning, vaccine-demand forecasting, antimicrobial stewardship and supply-chain contingency planning. These are potential applications, however, rather than demonstrated outcomes of the challenges themselves.

Eight challenges target major gaps in animal-health intelligence

1. Early warning and forecasting of animal disease outbreaks

The first challenge focuses on combining official disease notifications with other information sources, including syndromic signals, animal movements, trade activity and environmental data, to identify possible disease events earlier and estimate short-term trends.

Avian influenza and African swine fever are examples of priority diseases in this context. Both can have significant consequences for animal welfare, farm economics, veterinary interventions and international trade.

Traditional surveillance often depends on disease recognition, diagnostic confirmation and formal reporting. These processes remain essential, but delays between the emergence of an event and its notification can limit the time available for intervention. Integrating complementary datasets could help analysts identify unusual patterns that warrant further investigation.

For poultry producers, earlier warning of avian influenza risk could inform biosecurity reviews and preparedness. For pig producers, improved African swine fever intelligence could support contingency planning around animal movements, regional exposure and supply-chain disruption.

Any such system would still require validation against reliable surveillance data, clear treatment of uncertainty and expert interpretation. An algorithmic alert is not equivalent to a confirmed outbreak, and predictive performance will depend on the quality, coverage and timeliness of the underlying information.

2. AI-assisted extraction of disease signals from unstructured information

The second challenge addresses the difficulty of extracting useful animal-health information from news reports, technical publications, public notices and other unstructured sources.

AI-assisted systems could identify disease names, affected species, locations, event dates and other relevant details, then classify and organise them for further analysis. This could reduce the manual burden on analysts who monitor multiple countries and information channels.

For animal-health companies, the potential application is substantial. Regulatory affairs teams, veterinary technical services, market-intelligence departments and supply-chain planners all need timely information about disease events and changes in disease-control measures.

A structured intelligence feed could help users identify emerging issues and prioritise human review. However, automated extraction can misinterpret historical reports, confuse suspected and confirmed cases, duplicate the same event across publications or incorrectly associate a disease with a location.

Effective implementation therefore requires source traceability, deduplication, confidence scoring and human verification before important decisions are made.

3. Fairer comparisons of veterinary-system performance

The third challenge examines how countries’ veterinary systems can be compared more meaningfully when their resources, workforce availability and reporting capabilities differ.

A country with a larger veterinary workforce, more diagnostic laboratories or more developed reporting infrastructure may identify and document more disease events than a country with limited surveillance capacity. Raw counts alone can therefore provide an incomplete picture of underlying disease burden or system performance.

The challenge proposes adjusting indicators for differences in veterinary resources and national reporting capacity. Such analysis could help distinguish genuine epidemiological differences from variations caused by surveillance intensity or data availability.

For governments and international organisations, better-calibrated comparisons could support decisions about surveillance investment and technical assistance. For companies operating across several markets, they could provide additional context when interpreting reported disease patterns.

Adjusted indicators must nevertheless be designed carefully. Differences in geography, livestock populations, disease exposure, diagnostic access and reporting definitions can all affect comparisons. A statistical adjustment cannot compensate for every missing variable, and country rankings should not be treated as definitive measures of veterinary-system quality without supporting evidence.

4. Mapping trade networks and animal-disease vulnerability

The fourth challenge focuses on the relationship between livestock movements, trade connections, animal populations and disease events.

Trade networks can connect geographically distant production systems. Depending on the disease, the movement of live animals and other relevant pathways can influence exposure and the potential consequences of an outbreak. Mapping these relationships may help identify vulnerable connections and support scenario analysis.

For animal-health companies, this has direct relevance to business continuity. Disease-related movement restrictions, export suspensions and changes in regional disease status can affect the availability of livestock, veterinary medicines, vaccines, diagnostic supplies and animal-derived products.

Better-integrated trade and disease intelligence could help companies evaluate alternative supply routes, identify markets requiring additional monitoring and prepare for potential disruptions.

The challenge does not establish that a particular trade network is currently unsafe or quantify the economic effect of any future outbreak. Its value lies in the possibility of creating better analytical tools to assess risk before decisions must be made under pressure.

5. Identifying and monitoring animal-health misinformation

The fifth challenge addresses how misleading information about animal diseases emerges and spreads.

Misinformation can complicate outbreak communication, undermine confidence in vaccination and biosecurity measures, and make it more difficult for animal owners or livestock producers to distinguish reliable guidance from unsupported claims.

AI and data-analysis techniques could help researchers examine the circulation of disease narratives, identify recurring themes and monitor how misleading claims change over time. Such tools could assist veterinary authorities and professional organisations in identifying communication gaps and responding with clearer, evidence-based information.

For the industry, credible communication is important during disease outbreaks, product-safety events and vaccination campaigns. Veterinary pharmaceutical and diagnostic companies may also need to distinguish verified scientific evidence from claims circulating online.

Automated monitoring must be applied carefully. A widely shared claim is not necessarily false, and an unusual discussion is not proof of coordinated misinformation. Transparent methods, source evaluation and expert review are essential to avoid incorrectly classifying legitimate scientific debate or public concern.

6. Analysing long-term animal-disease trends

The sixth challenge proposes analysing historical WOAH disease data spanning 2005 through late 2024 to identify long-term patterns, unusual variations and possible associations with climatic or environmental factors.

Longitudinal analysis can help investigators examine how disease reporting changes across years and regions, whether patterns recur seasonally and which observations warrant further investigation.

Climate and environmental conditions may be relevant to certain vector-borne, wildlife-associated or environmentally influenced diseases. However, a statistical association does not by itself demonstrate causation. Changes in surveillance, diagnostic methods, reporting requirements, animal populations and disease-control policies can also influence observed trends.

For livestock producers, veterinarians and animal-health suppliers, a clearer understanding of historical patterns could improve risk assessment and inform future research priorities. For vaccine and diagnostic developers, it may help identify questions that deserve closer examination.

The historical dataset also has an important boundary: the challenge description identifies data through late 2024. It should not be described as a complete analysis of 2025–2026 disease events or as a real-time surveillance feed.

7. Forecasting antimicrobial use in animals through 2050

The seventh challenge focuses on the WOAH ANIMUSE database, which supports the monitoring of antimicrobial use in animals.

The proposed work includes analysing consumption relative to animal biomass, comparing participating countries and regions, examining antimicrobial classes and developing forecasts extending to 2050.

Antimicrobial-use data are relevant to veterinary medicine, livestock production, public health and antimicrobial stewardship. Understanding how consumption changes over time can help inform research and policy, provided that the data are interpreted in context.

For pharmaceutical manufacturers, distributors and veterinary professionals, improved analysis could support a clearer understanding of reported use patterns across species and markets. It could also help identify where additional information or stewardship interventions may be needed.

Consumption figures alone, however, do not establish whether medicines were prescribed appropriately, whether treatment outcomes improved or whether antimicrobial resistance increased. These questions require additional information, including indications for use, treatment practices, resistance surveillance and relevant animal-health outcomes.

Forecasts extending to 2050 should likewise be understood as scenarios based on assumptions and available data, not certain predictions of future antimicrobial consumption.

8. Building AI-assisted intelligence on veterinary antibiotic products

The eighth challenge proposes a structured repository of veterinary antibiotic products assembled from public national registers, regional databases and company product information.

The envisioned information includes product names, active ingredients, concentrations, pack sizes and target species. Organising these details into a consistent structure could help users navigate fragmented product information across markets.

For veterinary pharmaceutical companies and distributors, such a repository could potentially support regulatory intelligence, portfolio mapping, product comparisons and analysis of market availability. Researchers and veterinary authorities could also benefit from more accessible information on the products authorised or listed in different jurisdictions.

The challenge is particularly relevant to multinational operations, where product names, formulations, packaging and authorisation details may vary by country.

A usable database would need robust controls for data quality and currency. National registers may differ in format, update frequency and terminology; product authorisations can change, and similarly named products may have different formulations. Any resulting tool would need to preserve links to official sources and distinguish current authorisations from historical records.

The challenge does not itself constitute a new antibiotic approval, endorsement of a particular product or change to national prescribing rules.

Conclusion: a potential step towards more integrated animal-health intelligence

WOAH’s eight datathon challenges cover a broad range of priorities, from outbreak prediction and trade-network analysis to antimicrobial-use forecasting and veterinary antibiotic-product databases.

Together, they highlight the potential for data science to strengthen the connection between disease surveillance, veterinary decision-making and industry preparedness. For pharmaceutical manufacturers, vaccine developers, diagnostic providers, distributors and livestock producers, the most relevant opportunities lie in better information quality, earlier identification of risks and more consistent analysis across markets.

The next milestone is evidence of implementation: the analytical methods developed, the datasets used, their validation results and their practical value to users. Until those details emerge, the initiative should be viewed as a structured effort to identify and address important animal-health intelligence gaps—not as proof that AI has already transformed global disease surveillance.