Technology vs Politics: Critiquing Artificial Intelligence in Global Health

• In Event: Technology, Surveillance, and Artificial Intelligence in Global Health

Abstract

For decades, global health actors have heralded technological innovations as the cure to health inequities around the world. While these inequities have deep political determinants, global health actors have attempted to address them by depending on technological innovations ranging from medical technologies such as vaccines, drugs and diagnostics, to digital health technologies such as telehealth and electronic medical record systems. Today, artificial intelligence (AI) is emerging as the latest technology-based solution aimed at equalizing access to lifesaving healthcare, which has a variety of applications. A World Health Organization (WHO) report published in 2021 highlights emerging areas for AI that contribute to primary healthcare services: phone applications such as automated chatbots that would assist patients with self-care for non-communicable diseases, wearable sensors to monitor vitals and suggest “just-in-time” intervention, and AI-assisted algorithms that could prioritize where scarce primary healthcare resources would be best allocated. These and other forms of AI technology are becoming increasingly popular among global health actors.

Major bi- and multilateral agencies, including the Organization for Economic Cooperation (OECD), World Health Organization (WHO), and United States Agency for International Development (USAID), have convened forums and reports to advance the AI agenda in global health. There are even university-level courses taught on AI’s role in global health. Academic literature reflects these developments as well. A simple search of PubMed, a public health and medicine database, combining terms “artificial intelligence” and “global health” shows a marked rise in articles published on the topic; from eight in 2017 to 186 in 2022. The same search in the Worldwide Political Science Abstracts database reveals a similar trend, though with far fewer articles and a dropoff in articles on the topic in 2022 – indicating that there is ample space for further contributions to this debate from political science perspectives.

Before this prioritization of technological innovation, creating and supporting systems of primary healthcare existed at the center of global health policymaking. Indeed, the historic signing of the Alma Ata declaration inaugurated a new set of health goals for the world, proclaiming not just that health was a human right, but also that "governments have a responsibility for the health of their people," and that "primary health care is the key to attaining this target as part of development in the spirit of social justice" (UNICEF and World Health Organization 1978). These goals became known in the global health community as “Health for All.”

In this paper, we explore the ways in which expanding and prioritizing AI in primary healthcare risks further delaying action on and obscuring the political determinants of the Health for All agenda, which must be addressed to truly reduce health inequities. To do so, we conduct a critical appraisal of peer-reviewed and grey literature on AI in global health from 2017 onwards. Our ‘critical appraisal’ approach to the literature analysis seeks to critique, reveal, and challenge the power structures underpinning the rise, dynamics, and consequences of the AI agenda in global health.

While we do not dispute that AI in primary healthcare can play some role in improving health outcomes, based on this analysis, we caution that AI is the latest in a long line of technological innovations by powerful global health actors whose effects undermine efforts to strengthen primary healthcare. It does so by redirecting resources and priority from policies to strengthen health worker education, remuneration, and retention; assure adequate provision of medical supplies; and build and maintain infrastructure vital to functional primary healthcare systems, such as laboratories and generators. Ultimately, these processes uphold the structural violence of perpetually under-resourcing integral elements of primary healthcare systems.

Authors

• Lucia Vitale, University of California, Santa Cruz

• Leah Shipton