Belgium – IT services: consulting, software development, Internet and support – Ontwikkeling van een AI-gemedieerd kennissurveillancesysteem voor milieugezondheidszorg

The Environment and Health Team of the Department of Health is responsible for the preparation, execution, and substantiation of preventive health policy in the area of environmental health in Flanders. Environmental health is the component of preventive health policy that, through a multidisciplinary approach, with at least the scientific disciplines of medicine, toxicology, epidemiology, health care, environmental technology and environmental science, focuses on the prevention, early detection and intervention of environmental health damage. Within the Flemish Medical Environmental Network (VMMN), the team works with various governments, knowledge institutions and partner organizations. The core activities of the team include, among others, climate-health, risk analysis, promoting a healthy public space, environmental health focus areas, recreational and drinking water quality, indoor environment, nature-health and soil-health. Analyses are carried out to estimate the environmental-related health impact, and advice is formulated on environmental health aspects. However, the knowledge landscape within environmental science is growing particularly rapidly. New scientific insights, technological developments, changing legislation and international societal developments are creating a continuous stream of information. The systematic monitoring, assessment and application of this knowledge is currently largely manual and requires considerable personnel efforts. As a result, the risk of knowledge gaps increases, while the Department of Health is expected to base advice, policy decisions and health impact analyses on the most current and reliable insights. To address this challenge, the Department of Health wishes to develop an AI-mediated knowledge surveillance system. This system must automatically detect, process, structure and make relevant information from internal and external sources accessible to policy officers and experts. The solution supports the monitoring of scientific literature, policy documents, reports, legislation and other knowledge sources and translates these into usable document briefs and syntheses. The system focuses on the four central narratives within Environmental Health: Very Concerning Substances, Climate and Health, Healthy Public Space and One Health. Knowledge is ordered according to the source-exposure-dose-effect chain (BBBE), so that users can more quickly find and apply relevant insights in dossiers, risk analyses, policy preparation and advice. These initiatives have their legal basis in the preventive decree, the decree on general provisions for environmental policy with regard to the environmental hygiene aspects, the MER decree; and are implemented in accordance with the prioritisation of the health objective Environmental Health and the climate health plan. The envisaged solution supports human expertise but does not replace it. Interpretation, validation and policy guidance always remain the responsibility of substantive experts. Therefore, a human-in-the-loop approach is chosen, in which AI-generated output is supplemented with quality controls and human validation. The solution must also meet the current requirements regarding privacy, information security, AI governance and legislation, including the GDPR and the EU AI Act. It must connect to the Microsoft- and Windows-environment of the Department of Health and be coordinated with the Policy Information and Data (BID) and Communication and IT (CIT) departments, so that a secure, scalable and sustainably manageable solution is created that can also be applied in other teams within the Department of Health.