The rise of AI in LMICs: A new era for global healthcare

Source: Supplied. Dr. Ricardo Baptista Leite, CEO: HealthAI.

Without the constraints of established systems, LMICs have the flexibility to experiment with new technologies and models, offering a chance for truly transformative change. By leveraging AI to redesign health systems, these nations could challenge entrenched approaches in the global North. Fostering innovations in LMICs paves the way for more equitable healthcare solutions that benefit both local populations and the global community.

Primary care as the bedrock of a functioning health system

In order to facilitate transformation, it isn’t always necessary to reinvent the wheel— instead, the key lies in recognising and amplifying existing strengths. LMICs don’t need to outrightly replicate the hospital-heavy models of their wealthier counterparts. Rather, they can build upon their deeply rooted systems of community-based interventions.

Many LMICs have long relied on community health workers (CHWs) and decentralised primary care models to deliver essential health services, especially in rural and underserved areas. There lies great opportunity to improve access to quality healthcare by enhancing existing networks, training local health workers, and integrating AI-driven decision support, all without disrupting familiar systems that communities already place their trust in.

Community health workers, including —family doctors, nurses, and trained laypeople— often serve as the first point of contact for patients, even without specialised training. AI-driven recommendations can enable them to perform preliminary assessments with greater accuracy, quickly identifying those who require further evaluation or close monitoring.

AI can support CHWs by providing real-time decision support in diagnosing skin cancer through image analysis and detecting malaria from peripheral blood sample analysis, and as portable diagnostic tools like ultrasound probes and microscopes become more accessible, the range of AI applications in healthcare is expected to expand.

Additionally, with the growing use of smartphones, AI-powered tools can also offer ongoing support to individuals through lifestyle and nutrition guidance, enabling symptom self-assessment, and providing crucial advice during pregnancy and post-illness recovery, thereby easing the pressure on clinics and health systems.

#WEF24: Startups to test their AI models on European Commission supercomputers

Katja Hamilton  24 Jan 2024

It’s promising to see AI playing a growing role in improving patient management, from optimising electronic health records (EHRs) to enhancing clinical decision making in LMICs. In Malawi, where approximately 19 out of 1,000 newborns die during delivery or within their first month, the implementation of AI-enabled fetal monitoring at the Area 25 health centre has led to a significant 82% reduction in stillbirths and neonatal deaths at that facility.

By providing continuous, real-time monitoring of fetal vital signs during labour, this technology enables the early detection of fetal complications and facilitates timely interventions. This demonstrates how AI can harmoniously complement the human element—empowering, rather than replacing, the expertise of community health workers.

Although in countries like Mexico, where AI adoption in health remains slow, there are signs of optimism and initiative among physicians. According to a 2024 study, only 9% of Mexican doctors currently use AI-based tools in their daily practice, with nearly half preferring Chat-GPT like platforms and a quarter relying on specialised healthcare AI.

However, with proper training and institutional support, many more physicians are open to working with these technologies— over 53% of respondents view AI as a valuable tool for streamlining care and aiding complex decision making. Radiology, in particular, is emerging as a standout field where AI-powered imaging tools enhance diagnostic accuracy and efficiency by detecting subtle anomalies in X-rays and MRIs.

In short, LMICs have structural and societal advantages that make them particularly well suited for strengthening community-based healthcare and primary care systems. When these inherent strengths are paired with enabling technologies, transformation and innovation become the natural next steps. By integrating AI in a way that complements local strengths and needs, it also creates the ideal environment for building the trust necessary for communities to embrace it in healthcare.

A tangible readiness amid obvious challenges

From infrastructural deficiencies and financial constraints to workforce capacity issues, there are several challenges that impede the adoption of AI in healthcare in LMICs. Many of these countries lack the technological foundation necessary for implementing AI applications, while in many regions, the lack of robust data infrastructure stands in the way of collecting high quality, standardised health data for training and validating AI systems.

For example, in Pakistan many healthcare facilities still depend on outdated Electronic Health Record (EHR) systems and legacy technologies that are not designed to accommodate the evolving needs of modern healthcare.

Often lacking interoperability, this makes sharing patient information across various healthcare providers difficult, resulting in fragmented patient care. In countries like Lebanon, where ongoing financial crises, in all their severity, have led to critical shortages of even essential medicines, immediate healthcare needs naturally take precedence over the allocation of funds for advanced technologies like AI.

Public trust and acceptance also play a decisive role in AI adoption, with patients and healthcare providers often expressing skepticism due to concerns over accuracy, bias, safety, privacy, and job displacement.

Ogilvy One CTO Ravi Pal: Strategies for future-proofing AI and delivering customer-centric growth

1 day

Amid challenges, LMICs are increasingly demonstrating readiness for AI adoption in health through strategic policy development, investment in digital infrastructure, and innovative pilot programmes. Rwanda, for instance, became the first low-income country to introduce a national AI policy in 2023, outlining a structured approach to AI adoption across multiple sectors, including healthcare.

The policy emphasises ethical and inclusive AI use, positioning Rwanda as a forerunner in AI governance in the region. By investing in infrastructure, capacity building, and international partnerships, Rwanda is setting a solid foundation for the responsible and effective integration of AI in healthcare.

Despite being in its nascent stages and the absence of regulatory instruments specific to AI in health currently, LMICs including Rwanda, are proving that innovation can thrive in resource-constrained environments.

Other laws that are implicitly applicable to AI in health such as those governing data protection and privacy, cybersecurity, and information and communication technologies, play a vital role in shaping the responsible development and implementation of AI innovations in health. This, alongside a growing pool of tech interventions, mobile penetration and regulatory flexibility, creates a fertile ground for impactful, sustainable health systems to flourish.

Through the regulatory lens

In many LMICs, innovators face an unclear AI governance landscape, often leading to confusion or the false assumption that AI in health is unregulated. This uncertainty is further reinforced by the misconception that general AI regulations do not apply to the healthcare sector.

Many LMICs including Rwanda, Colombia, Lebanon and Pakistan are not without regulatory structures— they have existing systems that can be adapted to cover AI in healthcare more comprehensively.

Controversy erupts over R28m spent on NHI marketing

24 Feb 2025

Even in small or resource-limited countries, AI in health remains implicitly covered by laws meant for other medical technologies. For instance, many of these countries include software under their definition of medical devices, which extends to AI applications.

Are there gaps for general-purpose AI models through this regulatory approach? Yes. But can these existing frameworks serve as a foundational starting point for creating robust AI-specific regulations tailored locally for AI in health? Also, yes.

That said, ensuring the safe and responsible development of AI in health requires significant effort to update and refine global governance frameworks. Given the unique challenges LMICs face in developing regulatory mechanisms that address both local health needs and incorporate international standards, only a collaborative approach that balances global standards with the flexibility to adapt to local contexts can ensure equitable benefits to all populations.

Besides technical and ethical considerations, it’s also crucial to account for the cultural, historical, infrastructural and legal context of each country, as AI systems perform differently depending on these factors, reminding us why these nuances must not be overlooked.

An effective governance ecosystem can accelerate the development, deployment, equitable access, and adoption of AI technologies that substantially improve health outcomes. Given the right support, LMICs have the potential to drive locally grounded AI solutions and innovations in health— offering lessons for wealthier nations on equitable, sustainable and community-centric care.