Can artificial intelligence compensate for billions of dollars in disappearing global health funding?
Governments, philanthropies and technology companies are increasingly turning to AI as a potential way to address clinician shortages and expand access to care in low-resource settings. Technology companies and health systems continue to promote AI as a scalable solution for expanding access to care. But reporters covering these developments should be cautious about framing AI as a simple technological fix.
Indeed, studies show that AI research is overwhelmingly concentrated in countries with stronger health systems and fewer people living with certain diseases, raising concerns that the next generation of AI tools could reinforce instead of reduce global health inequities – making it crucial for journalists to ask tougher questions.
What’s the problem?
One of the central challenges to equitable AI deployment is what researchers call “health data poverty.” AI systems are only as good as the data used to train them. If those datasets mostly represent patients from high-income countries, the resulting algorithms may perform less accurately in populations that are underrepresented in the training data, including many communities in low- and middle-income countries and racial and ethnic minority populations elsewhere.
The second issue is who generates the evidence. AI development requires not only data but also research infrastructure, funding, technical expertise and regulatory pathways. Research suggests these resources remain concentrated in wealthier countries, potentially leading to earlier validation, deployment and commercial adoption in places where the clinical need may actually be lower.
That doesn’t mean AI will inevitably worsen disparities. India and China are examples of how sustained national investment in digital infrastructure and AI research can help shift innovation toward countries outside the traditional high-income research centers.
Questions to consider
When evaluating claims about AI in global health, look beyond technological performance and ask questions about implementation, representation and accountability.
Here are some questions to consider when reporting on AI in global health:
- Who was included in the datasets used to develop and validate this AI tool?
- Has the technology been independently evaluated in the populations where it is intended to be used?
- Who owns the data collected through global health AI partnerships?
- Is the research being led by investigators in the countries most affected by the disease, or primarily by institutions in high-income countries?
- Does the technology address a clearly identified health need, or is it being promoted as a substitute for investments in the workforce, public health infrastructure or health systems?
All of these questions get to the heart of the issue: whether AI is being developed and deployed in ways that genuinely expand access to care, or whether it risks leaving behind the very populations it promises to serve.
Additional resources
- World Health Organization – Global Initiative on AI for Health
- Johns Hopkins Bloomberg School of Public Health – Responsible AI in Global Health
- Harvard University – Research AI for Global Health and Public Health
