WHO Publishes Discussion Paper on AI and Evidence-Informed Health Policy — Calls for Algorithmic Impact Assessments and Human-in-the-Loop Oversight

The World Health Organization published 'Artificial intelligence and evidence-informed policy – emerging challenges and opportunities' on June 2, 2026, formally redirecting the UN AI-in-health policy agenda beyond clinical care to the foundational question of how AI reshapes the evidentiary basis of health policymaking itself. The paper calls for mandatory algorithmic impact assessments before deployment and human-in-the-loop oversight post-deployment across all AI health applications. It introduces the concept of 'epistemic injustice' — warning that AI systems trained on mainstream health data may systematically deprioritize Indigenous knowledge systems, traditional medicine practices, and lived patient experience in favor of quantifiable biomedical metrics. The paper identifies five emerging governance challenges: data quality and provenance, algorithmic transparency, accountability frameworks, equitable access, and the risk of AI-generated evidence creating new barriers for community-level health decision-making. This marks the WHO's first comprehensive treatment of AI's impact not just on clinical AI tools but on the underlying research and policy infrastructure that shapes global health decisions.