AI in Healthcare for Resource Limited Settings: An Exploration and Ethical Evaluation

Advances in artificial intelligence (AI) hold great promise to transform healthcare in resource-limited settings (RLS). However, due to challenges such as shortages in healthcare professionals, data scarcity, and inadequate regulatory frameworks, RLS are left especially vulnerable to AI's potential risks and ethical violations. Thus, despite rising expectations, substantial gaps remain in our understanding of how to responsibly integrate AI into the healthcare systems of RLS. In response, this review critically examines AI healthcare applications in RLS, with the intention of promoting ethical, transparent, and secure principles for future implementations. We first provide an exploration of the potential uses of AI in resource-limited healthcare and present four broad subfields: decision support, predictive analytics, telemedicine and digital health tools, and resource management. Taking an analytical approach, we illustrate both the potential benefits and hidden ethical pitfalls that may arise when implementing AI in contexts with limited human and financial resources. Drawing on both recent studies and original perspectives, we aim to provide an overview of major ethical concerns in an RLS context - including algorithmic bias, non-maleficence, privacy and security, autonomy, and transparency - as well as discussing additional ethical dilemmas rarely addressed in literature. Subsequently, we advocate for context-specific regulations and culturally sensitive frameworks, in addition to robust oversight and active local participation. Finally, we provide recommendations that aim to protect patient welfare, uphold autonomy, and promote equity-so that AI applications ultimately strengthen, rather than undermine, global efforts to reduce healthcare disparities.