Submission ID 131911
| Issue/Objective | AI is rapidly reshaping global health - yet the governance frameworks determining how it is developed and regulated are shaped overwhelmingly by high-income country governments, corporations, and institutions. Low- and middle-income countries (LMICs), which carry the greatest disease burden and stand to benefit most, remain peripheral in standard-setting processes. This structural exclusion risks entrenching power asymmetries under a technological veneer, producing systems that reflect high-income priorities while failing populations in the Global South. This presentation critically examines three landmark AI governance frameworks and proposes pathways toward equitable, inclusive governance. |
|---|---|
| Methodology/Approach | This review employs a critical policy analysis of three landmark AI governance instruments: the G7 Hiroshima AI Process (2023), the African Union's Continental AI Strategy (2024), and the WHO Ethics and Governance of AI for Health Guidelines (2021, updated 2024). Each framework is analyzed using a power-mapping lens assessing: LMIC representation in deliberative processes, alignment of governance principles with LMIC health system realities, and whose values and priorities are centered in normative language. Document analysis is triangulated with a review of peer-reviewed literature and grey policy sources published between 2021 and 2025. |
| Results | Across all three frameworks, significant disparities in agenda-setting power emerge. The G7 Hiroshima Process, though influential in establishing interoperability standards, was developed without formal LMIC representation - its principles reflecting high-income contexts with assumptions of robust data infrastructure and regulatory capacity largely absent in the Global South. The WHO guidelines articulate equity as a core principle yet cite fewer than 15% of evidence from LMIC settings, undermining contextual validity. By contrast, the AU Continental AI Strategy foregrounds African priorities - including data sovereignty, local language representation, and development-centred ethics - yet remains severely under-resourced and fragmented relative to G7-backed frameworks. |
| Discussion/Conclusion | These findings expose AI governance in global health as a site of power reproduction rather than redistribution. Without structural reform - guaranteed LMIC co-leadership in multilateral bodies, Southern-led research mandates, and context-sensitive regulatory capacity-building - AI risks deepening the inequities it claims to address. Governing AI equitably demands the solidarity and decolonial intent that the global health equity movement has always required. |
| Presenters and Affiliations | Uzochukwu Chima University of Nigeria Somtochukwu Anierobi University of Nigeria Chinonyelum Agbo University of Nigeria Maureen Ani University of Nigeria Amauche Ngige University of Nigeria |
