Two-thirds of the 53 WHO Europe countries already deploy AI diagnostics, and half have AI-powered patient chatbots. Yet only one in 12 has a strategy for governing AI. Hans Henri P. Kluge, the WHO Europe chief, warned that this mismatch puts patients at risk. Technology is reaching hospitals faster than public oversight can keep pace.
Deployment Before Governance
Kluge called the situation “concerning” and said it could take a toll on human health. “All of this erodes public trust in health systems more broadly,” he warned. His figures describe a gap between systems already in clinical use and the national strategies meant to govern them. Patients and healthcare professionals need confidence that AI is safe and its benefits outweigh its risks.
Approval isn’t the end of oversight. It must continue after a system enters clinical practice, when emerging risks and adverse events become visible. Governance has to extend beyond the promises made before deployment.
Prevention Needs More Than Software
Ricardo Baptista Leite, CEO of HealthAI, the global agency for responsible AI, said AI could help healthcare providers identify people at risk of certain conditions before they become ill. By analysing health data alongside social factors that influence health, the technology could help identify people who need support before their conditions deteriorate.
But tools don’t remake the systems that use them. “If you retrofit AI into an inefficient system, it just makes you very efficient at being inefficient,” Baptista Leite told Euronews. He added: “Technology is not going to save us alone.” His warnings point to a basic question about what healthcare systems reward: hospitals are often paid to treat patients rather than keep them healthy, he argued.
Europe’s ageing population is putting increasing pressure on healthcare systems already stretched thin. Baptista Leite said a move toward prevention would require new funding models that reward early intervention and proactive monitoring. Without that change, AI may identify a need earlier while the system’s incentives still point toward treatment after illness. The goal, he said, isn’t simply to save someone after a heart attack, but to prevent it from happening.
Who Watches the Systems in Use?
HealthAI works with regulators in 10 countries to develop a global early warning system. It aims to identify adverse events and share information about emerging risks. Baptista Leite said these mechanisms could help detect harm in time and keep citizens safe if something goes wrong.
The work addresses the oversight gap Kluge described: AI tools already operate across European hospitals, while only a few countries govern them. The figures don’t say which systems will cause harm or when. They do show how widely deployment has spread compared with national strategies. HealthAI’s proposed system aims to help regulators spot problems across borders, but it can’t replace continuing oversight once AI enters clinical practice.
Patients face more than the question of whether software can flag a risk. Health systems must decide how to fund prevention, monitor tools after approval and respond when adverse events emerge. The technology may help providers act earlier. It can’t make those institutional decisions for them.