Skip to main content

Ethics & Governance

Trust is a feature, not an afterthought. GeoAI4EI embeds ethical and legal safeguards into the methodology of every tool, so that AI-driven public health decision support can be relied on by agencies, policymakers, and the public.

Legal & regulatory compliance​

  • GDPR — all personal data handling complies with the General Data Protection Regulation
  • Data Sharing Agreements — formal agreements with every data provider
  • Anonymisation standards — data is anonymised before processing wherever required
  • EU AI Act — developed AI tools conform to the European AI Act's obligations for their risk category

Principles: FAIR & CARE​

  • FAIR (Findable, Accessible, Interoperable, Reusable) guides the data and software infrastructure
  • CARE (Collectively benefit, Are accountable, Respect knowledge, Equity) guides engagement with communities and data providers

Fairness-by-design​

Fairness is engineered into the pipelines rather than checked afterwards:

  • Bias audits of models and training data
  • Sex/gender covariates and analyses to detect and correct differential impacts
  • Systematic evaluation against the populations the tools will serve

Transparency​

  • Model cards document the purpose, data, limitations, and performance of each model
  • Uncertainty visualisation shows the confidence behind every estimate
  • Traceable sources for every answer produced by conversational assistants

Safeguards against over-reliance & misuse​

  • Clear positioning of all tools as decision support, never decision replacement
  • Guardrails on the capabilities exposed to end users
  • Monitoring of usage patterns and documented human-in-the-loop requirements

Digital inclusion​

  • Open-source tools that any health system can adopt
  • Lightweight platforms that run in resource-limited contexts
  • Training programmes for practitioners outside well-resourced institutions

Cybersecurity​

  • EU-based secure hosting for all project data and services
  • Encryption in transit and at rest
  • Role-based access control with the principle of least privilege

Independent audits & interoperability​

  • All AI tools undergo independent ethical audits
  • The project ensures interoperability with EU infrastructures (EHDS, ATHINA)
  • By embedding ethics and inclusiveness, GeoAI4EI enhances trust in AI-driven public health tools and reduces inequalities in pandemic preparedness
tip

Want to see how these principles show up in practice? Read the introducing GeoAI4EI blog post.