Key points are not available for this paper at this time.
Advances in artificial intelligence (AI) and earth observation (EO) have led to promising new opportunities for promoting sustainable development in the Global South. However, adapting AI for EO (AI4EO) solutions to real-world problems in these contexts are not without risks and challenges. In this paper, we discuss the barriers to successful adoption of AI4EO in developing countries and propose a roadmap for more impactful applications of future AI4EO research. The recommendations emerge from three case studies that illustrate the need to create AI4EO systems that take into account technical, social, and ethical considerations specific to the Global South such as data and infrastructural limitations, gaps in technological literacy, geodiversity issues in remote sensing (RS) datasets, and the need for stronger stakeholder engagement, among others. Finally, we propose a framework based on more inclusive, human-centered methodologies rooted in the needs, perspectives, and lived experiences of domain experts, decision-makers, and local communities in the Global South.
Tingzon et al. (Wed,) studied this question.