AI at the World Health Assembly 77: What are people talking about and what’s missing?
This year’s World Health Assembly (77th) coincided with the 2024 AI for Good Summit, marking a timely convergence as discussions at both events underscored the transformative potential of digital systems and artificial intelligence (AI) in health. Miyu Niwa and Tyler Smith represented Cooper/Smith at a number of key fora and roundtable events, including the World Summit on the Information Society (WSIS) High-Level meetings, Geneva Digital Health Day, and the AI for Good Summit.
What’s everyone talking about?
AI: Unsurprisingly, you couldn’t turn a corner without hearing the ubiquitous letters floating in the air. There is great hope for emerging new technologies to fill service gaps and support strapped economies to meet population needs. While the robots walking the floor at AI for Good are always a fun encounter, pragmatic innovations on the policy front are more likely to support low- and middle-income countries (LMICs) adopt useful AI tools. It seems the tone of the conversations has shifted from last year as well. Existential arguments about the ‘rise of robots’ have receded marginally, replaced with more substantive conversations about the differential benefits and costs AI will confer. We particularly liked the keynote with Geoffrey Hinton, where he talked about the very real threats of exacerbating economic inequalities and fomenting political instability and some unique ideas about combating them. Re: health specifically, global leaders seem bullish on the ability of AI to support the stretched health workforce.
DPI: Another new(ish) abbreviation on everyone’s lips refers to digital public infrastructure — the foundational building blocks and ecosystem that allow information to be easily digitized, shared, and accessed. This is a no-brainer and something we’ve been working on our whole careers in the health sector. The DPI approach and frameworks provide a rational, multi-sector blueprint for investment that prioritizes efficiency and economies of scale — 2 things that will make or break data use and the routinization of AI in LMICs. Many groups are focused on DPI for health, but it seems the WHO Global Initiative on Digital Health (GIDH) has thankfully taken the reins to coordinate.
SGDs: The world is not on track to meet the Sustainable Development Goals. Only 16% of SDG indicators are showing progress, stalled of course by COVID-19, food shortages, and global inflation. The COVID effect is also clear in the latest Global Burden of Disease Estimates, launched the same week. There’s a palpable and somewhat nervous hope that emergent technologies like AI will be the jolt in creativity and capacity we need to correct course. We’re cautiously optimistic.
What’s missing?
Frank and realistic discussion about infrastructure policies: AI requires substantial hardware, energy consumption, and computational power to work. In the majority of LMICs, data hosting and protection policy requires on-premise data storage and processing. While it is understandable that countries in the wake of colonial history would opt for maximum control, the technical and environmental requirements to build and maintain this infrastructure are largely out of reach and unsustainable. If we can’t figure out a solution to meet the privacy and data ownership needs of LMICs, while pooling expensive and highly sophisticated technical assets, these countries will be left far behind in the AI future.
Useful tools to measure the incremental value of tech: Every country’s government has to make trade-offs. In LMICs, trade-offs across sectors often result in underfunded health programs. Within health, these trade-offs often mean deciding who goes without much-needed services, drugs, or basic utilities. Yes, AI for health can be transformative, but adoption will require convincing the funders, the clinic managers, the doctors, and the nurses, that a digital intervention is (a) valuable and (b) superior to what they already do. In a place where stock-outs for essential drugs are the norm, how would you convince a funder that $100,000 is better spent on an algorithm to reduce drug shortage, rather than buying those missing meds? We lack standard, useful tools to properly evaluate new digital tech in context. Our colleagues at the World Bank are certainly thinking about this and luckily raised these issues during the week. We need some serious advancement in our methods and tools in this area.
A focus on reliability and trustworthiness in simple use cases: With big hopes, comes big expectations. The excitement around AI for health is very much skewed toward apps that support health providers and patients directly (think of AI vision to better read diagnostics and chatbots for medical advice). While we are all for this future, there are some very real challenges and tangible benefits to deploying AI tools in support of users more upstream in the health system. Large language models may not be able to provide reliable medical advice (don’t eat rocks!), but they are very good at parsing large troves of unstructured info to find what you’re looking for. They’re also good at coding. We see some very clear use cases that can help make sense of complicated clinical guidelines and health policy, parse mountains of health stats to return that needle in the haystack, and generally meet non-technical users where they are. Implementing these use cases requires some advances in structure and reliability, things we are working on now at Cooper/Smith.
The week proved to be a pivotal moment in global health, highlighting the need for collective global efforts to ensure clear guardrails and equitable approaches to AI use. At Cooper/Smith, we are undoubtedly excited about the transformative power of AI in global health, especially for healthcare planning and decision-making in LMICs to enhance health service delivery, promote equity, and improve health outcomes. We are eager to continue shaping discussions on responsible AI use tailored to the unique demands of global health.
