WHO report warns against overestimating benefits of AI in health

AI can help to bridge the health service access gap in poor countries, but bias, unethical use of data, and poorly trained models are all risks to be overcome

The World Health Organisation (WHO) has said that AI holds great promise for improving healthcare, but only if human rights and ethics are put at the heart of its design and use.

The WHO has created a new report on, titled 'Ethics and governance of artificial intelligence for health' (pdf), following two years of consultation with a panel of international experts. It details various opportunities that AI technology offers to communities and governments in healthcare systems, but also warns about unregulated use.

"Like all new technology, artificial intelligence holds enormous potential for improving the health of millions of people around the world, but like all technology it can also be misused and cause harm," said WHO Director-General Dr Tedros Adhanom Ghebreyesus.

"This important new report provides a valuable guide for countries on how to maximise the benefits of AI, while minimising its risks and avoiding its pitfalls."

Several countries are already using AI to improve the accuracy of disease diagnosis; support drug development and health research; assist with clinical care; and improve diverse public health interventions, like disease surveillance and health systems management.

The WHO says AI could empower patients to take greater control of their own healthcare. It may also enable poor countries to bridge health service access gaps.

However, it also cautions that AI opportunities are also linked to risks, including bias in algorithms, unethical use of health data and threats to patient safety and the environment.

Moreover, it points out that AI models trained primarily on data collected from people in richer countries may not give similar results for individuals in middle- and low-income settings. Poor training data like this is a recurring issue when it comes to AI systems.

To regulate the use of AI, the WHO has published the following six guiding principles, which it believes will help to limit risks in AI design and use, while maximising opportunities for governments and communities.

The principles are:

The WHO report comes a couple of weeks after the UK government announced £36 million funding for AI projects, with the aim of helping to revolutionise diagnostics and care in the NHS.

The government said that the fund will go to 38 new AI research projects, selected as the winners of the second round of £140 million funding under the NHS AI Lab's 'AI in Health and Care Award' programme.

The projects chosen to receive funding include two cancer detection tools; a chatbot app to help identify anxiety and depression in patients; a new approach to identifying spinal fractures in CT scans; and an AI-guided tool to help doctors diagnose heart attacks more accurately.