OECD health ministers: people and partnerships are the secret to enabling AI in healthcare

AI in healthcare illustration

Health systems are facing multiple crises at the same time – increasing costs and demands due to an ageing population, a need to strengthen public health to be ready for future health crises, and a projected workforce shortage of 3.5 million by 2030 across the OECD  (OECD, 2023[9]). Health leaders are looking for solutions and hope.

Artificial intelligence (AI) has provided that hope. AI algorithms are saving lives by assisting healthcare providers to detect hard-to-see anomalies and allow preventative action. AI can automate administrative or routine functions to give health providers more time to provide care (Beamtree Global Impact Committee, 2023[5]). Patients communicating with AI systems are having more empathetic and understandable interactions (Tu et al., 2024[3]). AI is also making identifying and testing new antibiotics faster (Liu et al., 2023[60]). New stories about AI’s opportunities in health are published daily.

Follow us on LinkedIn

Challenges to AI in healthcare

There are also cautionary tales about the role of AI in health, including concerns about the privacy of sensitive personal health information, bias from AI solutions trained on poor quality data, and the ethics and liability of using apparent black boxes in the provision of care.

Another concern with the introduction of AI in health systems is growing inequities. Relatively wealthy health institutions invest in health-related AI and offer those solutions to those who can afford it. As such, only a privileged few may enjoy the benefits of AI, widening divides.

All those risks must be tackled. Yet, this should not prevent health systems from actively looking at leveraging AI to deliver better care and support research.  Failure to actively drive the future of AI in health risks bespoke solutions that are fragmented and biased. This has happened before with digitalisation in health systems, where fragmentation of health data led to difficulties connecting hospitals, primary care, and specialists. COVID-19 showed that countries with stronger data foundations were able to manage their response to the pandemic better. It is crucial that we learn the lessons from the last 20 years and engage with AI in health in a way that puts people first and leaves no one behind.

Enabling health data in AI and guiding principles

For several years, the OECD has worked with countries to encourage using health data that respects privacy and promotes innovative and trustworthy AI. A 2022 report that reviewed how countries progressed against the 2016 OECD Council Recommendation on Health Data Governance showed how many countries had difficulty integrating data to improve health outcomes and health system performance. For AI, the OECD developed principles in 2019. These principles emphasise the importance of human-centred values, inclusion, transparency, safety, and accountability.

It is time to turn these principles and recommendations into action – to ensure that AI in health is developed responsibly and that the AI solutions are trained on quality health data while delivering timely solutions for better health outcomes and protections for all. Such action would establish a patient-centred policy, data, and technical environment for adopting responsible AI in health solutions.


Advancing Responsible AI requires agility, partnerships, and sound governance principles

The OECD published Collective Action for Responsible AI in Health on 22 January 2024, emphasising action to enable responsible AI in health that would improve trustworthiness, build capacity, evaluate problems, bring solutions, and accelerate progress. This was echoed in the OECD Health Ministers’ Declaration on Building Better Policies for More Resilient Health Systems that came out of their meeting on 23 January 2024. The OECD report presents four recommendations for collective action:

  • Adopt a nimble approach to regulating and overseeing AI in health: Legal and regulatory frameworks should be modernised to address the unique challenges posed by digital solutions and AI in health. Regulations would create new standards for AI-related to safety, quality, transparency, economic evaluation, liability, and accountability while updating regulations for health data access, privacy, and security that protect citizens. Such regulation would recognise the rapid development of innovation and establish proactive and responsive oversight.
  • Establish public-private and person-provider-partnerships: There are opportunities to encourage relationships between the public and private sectors and people impacted by AI in health solutions. This would ensure that protections are in place for public health data rights while leveraging the complete set of capabilities to use health data and AI for the public good. It would also ensure that solutions are people-centred and driven by their needs.
  • Elevate health data governance: Health data governance should ensure that quality data is available for AI innovations. Strengthening health data governance is one of the most impactful ways to mitigate the risk of cyber threats while lowering the cost and complexity of innovation. For example, interoperability across health institutions that simplifies data collection for AI and enables AI solutions to scale up should be championed – while continuing to ensure that appropriate health data protections are in place.
  • International collaboration and leadership for greater results: Collaboration across countries and through international organisations helps to transfer best practices and share knowledge, accelerating progress and leading to better and safer AI benefits. The collaboration will also help develop, spread, and scale up innovation within and across borders.

The right time to focus policy on AI in health is now. The opportunities to generate value through AI solutions and bend the cost curve by using health data in a better way are tremendous. To do this, we must create the right environment to deliver value for all.

Disclaimer: The opinions expressed and arguments employed herein are solely those of the authors and do not necessarily reflect the official views of the OECD or its member countries. The Organisation cannot be held responsible for possible violations of copyright resulting from the posting of any written material on this website/blog.