The gap the conference was built to address The meeting was triggered by a specific, uncomfortable number. WHO/Europe's own assessment the most comprehensive review of AI readiness it has ever conducted across the region found that nearly two-thirds of countries in the WHO European Region are already deploying AI in clinical diagnostics, while only 8% have a health-specific AI strategy, and only 8% have liability standards defining who is responsible when an AI system fails. "Let me start with a number: 8%," WHO Regional Director for Europe Dr. Hans Henri P. Kluge told delegates on July 15, opening with the governance figure rather than the adoption one. He framed the gap between how fast AI is being deployed and how slowly it is being governed as "the defining challenge of AI in health right now." Later in the conference, he put the urgency more bluntly: regulating AI in health is hard, he said, but not regulating it is harder measured in patients harmed, trust lost, and inequalities widened. Three pillars, not one framework Rather than proposing a single global standard, the conference organized its work around three pillars: the rules that govern how AI is regulated and held accountable, the tools and infrastructure needed to deploy it safely, and the people and institutions responsible for putting it into practice. Discussions spanned legal accountability, data governance, interoperability, workforce preparation, and responsible investment explicitly designed to be adaptable across countries at very different levels of digital development, rather than a single model imposed uniformly. Equity ran through all three pillars. A central concern raised repeatedly in Lisbon was whether AI in health will narrow the gap between well-resourced and under-resourced health systems, or widen it the same tension that shapes how African institutions have approached AI adoption on their own terms. The one concrete deliverable: a Lusophone roadmap If Lisbon produced a single tangible output, it was regional rather than global. A dedicated working session brought together Angola, Brazil, Cabo Verde, Portugal, São Tomé and Príncipe, and Timor-Leste the Portuguese-speaking bloc to lay groundwork for a joint cooperation framework on AI and health among Lusophone countries. It is a narrow instrument, built for one language community, not a template the rest of the world is expected to adopt. Where this leaves Africa Here is what distinguishes Lisbon from a genuinely global governance moment: Africa was not the driver of this process, and the continent has already been building its own answer on a separate track. Africa CDC has led development of a Continental Health Data Governance Framework, presented for African Union endorsement at the AU Summit in February 2026 — work that predates Lisbon and sits alongside a broader Continental Strategy on AI that the AU endorsed back in July 2024. That framework is meant to harmonize policies on health and genomic data across AU member states, establishing shared standards for privacy, consent, ownership, and cross-border data sharing, with an explicit goal of reducing vendor lock-in and external dependence rather than importing governance models built for a different set of constraints. That distinction matters for how African newsrooms should read Lisbon. WHO/Europe's conference was convened to solve a European readiness gap that happens to have global echoes, not to set continental policy for Africa. Whether and how Africa CDC's own framework engages with whatever working agenda emerges from Lisbon on interoperability standards, liability norms, or workforce training is a genuinely open question, and one AHM will continue tracking as both processes move forward in parallel. What's next WHO/Europe has said it will continue advocating that AI regulation in health is necessary, not optional, as the technology becomes embedded in how care is delivered. No text of a binding agreement emerged from Lisbon, and none was promised going in delegates left, in Kluge's own framing, with a working agenda rather than a communiqué. The real test of the conference's value will be whether that working agenda translates into liability standards and national strategies in the countries that currently have neither. Photo: Delegates at the Global WHO AI Conference: Shaping AI in Health, held 15–16 July 2026 in Lisbon, Portugal. Source: WHO/Europe. _Reporting sources: WHO/Europe press release (15 July 2026); WHO/Europe AI readiness assessment; The Lancet Regional Health – Africa; AUDA-NEPAD._