The World Health Organization (WHO) has released a new discussion paper, Artificial intelligence and evidence-informed policy: emerging challenges and opportunities, examining how AI is starting to reshape the way health policies are designed, implemented and monitored around the world.

The paper looks at “evidence-informed policy-making” (EIP), the process through which research findings, data and practical experience are brought together to design, apply and improve public health policies. According to WHO, AI can support this process at every stage: it can help integrate large and varied datasets, speed up the analysis of scientific evidence, model different policy scenarios, and provide continuous feedback as policies are implemented.

However, the report points out that these possibilities come with challenges. WHO highlights several cross-cutting risks that can affect any phase of the policy cycle, including the difficulty of understanding how some AI systems reach their conclusions, the concentration of AI development in the hands of a few large companies, regulatory gaps, and the risk that public officials may come to rely too heavily on automated tools at the expense of their own judgement. The paper also stresses that AI systems trained on incomplete or biased data can reproduce or even worsen existing inequalities.

Rather than proposing entirely new rules, WHO’s paper suggests building on frameworks that already exist in both the health-policy and AI-governance worlds, from data-quality standards to risk-based regulatory approaches such as those found in the EU AI Act. Among its practical recommendations, the paper calls for keeping humans firmly at the centre of decision-making, encouraging multidisciplinary teams to review AI outputs, running regular bias and impact assessments, and using pilot projects to test tools before they are used at scale.

Overall, the paper is intended as a starting point for discussion rather than a definitive guide. It calls for closer collaboration between governments, international organisations, researchers and civil society to ensure that AI is used to strengthen evidence-informed policy-making.

The full discussion paper is available on the WHO website.