Artificial Intelligence in Medicine And Public Health
1. Kalybekova K.D.
2. Hainal Neha
Rathod Adarsh
Mumtaz Masheera
(1. Teacher, Public Health Dept., International Medical Faculty, Osh State University, Osh, Kyrgyzstan.
2. Students, International Medical Faculty, Osh State University, Osh, Kyrgyzstan.)
Abstract
Artificial intelligence (AI) is rapidly transforming medicine and public health at a global scale, offering unprecedented opportunities to improve diagnostic accuracy, accelerate drug discovery, strengthen disease surveillance, and enhance health equity. This research article synthesizes evidence from authoritative international sources — including the World Health Organization (WHO), PubMed/NCBI, UNICEF, the United States Food and Drug Administration (FDA), and leading peer-reviewed journals — to provide a comprehensive examination of AI's current applications, demonstrated benefits, and unresolved challenges in clinical medicine and public health. The article reviews AI's role in diagnostic imaging, pathology, clinical decision support, pharmaceutical development, epidemiological surveillance, maternal and child health, and pandemic preparedness. It also critically examines the ethical framework established by WHO in 2021, issues of algorithmic bias and health equity, and the particular challenges facing low- and middle-income countries (LMICs). The article concludes with a forward-looking discussion of governance requirements, infrastructure needs, and global collaboration strategies necessary to ensure that AI serves as a force for universal health benefit rather than a driver of new inequities.
Keywords: Artificial Intelligence, Machine Learning, Deep Learning, Public Health, Clinical Medicine, WHO, Disease Surveillance, Diagnostic Imaging, Drug Discovery, Health Equity, Algorithmic Bias, Low- and Middle-Income Countries, COVID-19, Pandemic Preparedness, Digital Health.
1. INTRODUCTION
The convergence of exponential growth in computational power, the unprecedented availability of health-related data, and transformative advances in machine learning (ML) and deep learning (DL) algorithms has positioned artificial intelligence as one of the most consequential innovations in the history of medicine. From analysing medical images with radiologist-level precision to forecasting disease outbreaks weeks before traditional surveillance systems would detect them, AI is redefining what is clinically and epidemiologically possible.
The World Health Organization (WHO) has unambiguously acknowledged AI's transformative potential, stating that it 'can be, and in some wealthy countries is already being used to improve the speed and accuracy of diagnosis and screening for diseases; to assist with clinical care; strengthen health research and drug development, and support diverse public health interventions, such as disease surveillance, outbreak response, and health systems management' (WHO, 2021). The WHO Director-General, Dr. Tedros Adhanom Ghebreyesus, has further emphasised that AI 'holds enormous potential for improving the health of millions of people around the world,' while cautioning that it 'can also be misused and cause harm' (WHO, 2021).
Simultaneously, the scale of global unmet health need makes the stakes extraordinarily high. According to UNICEF's 2024 Levels and Trends in Child Mortality Report, despite decades of progress, millions of children still die from preventable causes each year, with the most vulnerable concentrated in sub-Saharan Africa and South Asia. UNICEF's 2024 Digital Health and Information Systems Annual Report confirms that digital solutions — including AI-powered tools — are being increasingly integrated into national health systems to improve service delivery and data-driven decision-making, particularly in fragile settings.
This article provides a rigorous, evidence-based synthesis of AI's role in transforming medicine and public health. It draws upon WHO policy documents, systematic reviews and meta-analyses indexed on PubMed, UNICEF data, FDA regulatory guidance, and the most current peer-reviewed literature to present a structured account of where AI is succeeding, where it is falling short, and what must be done to realise its full potential equitably and responsibly.
2. ARTIFICIAL INTELLIGENCE IN CLINICAL MEDICINE
2.1 Diagnostic Imaging and Radiology
Radiology represents the most advanced and best-evidenced domain of AI application in clinical medicine. AI-driven technologies — particularly convolutional neural networks (CNNs) and deep learning architectures — have demonstrated the ability to interpret medical images including X-rays, computed tomography (CT) scans, magnetic resonance imaging (MRI), ultrasound, and digital pathology slides with accuracy that meets or exceeds that of experienced clinicians in specific tasks.
A 2024 systematic review published in Cureus (Strubchevska et al., 2024, NCBI PMC11582495) examined eight peer-reviewed studies published between 2018 and 2024 and concluded that AI is transforming the diagnostic methods of radiology by automating image interpretation, improving diagnostic accuracy, and optimising clinical workflow. Key findings include:
AI deep learning models have demonstrated high sensitivity and specificity in detecting pulmonary nodules on CT scans, early-stage breast cancer on mammography, and ischaemic changes on brain MRI, often identifying subtle findings missed in initial reads.
The CAMBNET model, evaluated on 160 invasive breast cancer cases using dynamic contrast-enhanced MRI, achieved diagnostic accuracy of 88.44% and an Area Under the Curve (AUC) of 96.10%, with an AUC of 99.95% for a specific subtype classification (Frontiers in Radiology, 2025).
AI algorithms significantly reduce radiologist workload by automating routine tasks such as image segmentation, measurement, and preliminary reporting, allowing specialists to focus on complex diagnostic challenges.
A parallel 2024 article in Cureus (Bhandari, 2024, PMC11521355) underscored that AI-driven technologies — including machine learning, deep learning, and natural language processing (NLP) — are 'playing a pivotal role in automating routine tasks, aiding in early disease detection, and supporting clinical decision-making' across radiology subspecialties.
2.2 Pathology and Histopathological Analysis
Digital pathology, powered by AI, is another rapidly maturing field. Histopathological examination of tissue specimens remains the gold standard for definitive cancer diagnosis, and AI is now augmenting — and in certain supervised contexts approaching replacing — portions of this process.
A 2025 review published in Military Medical Research (Springer Nature) notes that 'AI offers transformative potential in pathology, where histopathological images remain the diagnostic gold standard due to their rich morphological and molecular information.' The review provides a technical taxonomy of AI algorithms applied across the diagnostic continuum — from image preprocessing and tumour classification to prognostic stratification and biomarker discovery.
The International Agency for Research on Cancer (IARC), part of WHO, reported 19.29 million new cancer cases and 9.96 million cancer-related deaths globally in 2023 (Global Cancer Burden Data, 2023). The burden of cancer, combined with severe shortages of pathologists in many countries, makes AI-assisted pathology tools particularly promising as a means of scaling cancer diagnostic capacity without proportionate increases in specialist workforce.
2.3 Clinical Decision Support Systems
Beyond imaging, AI-powered clinical decision support systems (CDSS) are being deployed to assist clinicians in diagnosis, risk stratification, treatment planning, and medication safety. Natural language processing enables AI to extract clinically relevant information from unstructured electronic health records (EHRs), discharge summaries, and clinical notes, dramatically reducing the administrative burden on healthcare providers.
A 2025 comprehensive review published in PMC (PMC12468291) examined AI applications across diagnostic imaging, clinical decision support, surgery, pathology, and drug discovery using PubMed as the primary database. The review confirms that 'as of 2025, only a limited number of AI medical devices have regulatory approval (FDA-cleared algorithms, mostly in radiology and cardiology, for specific detections),' while highlighting that the regulatory pathway involves demonstrating safety and effectiveness — a significant ongoing challenge for dynamic AI systems.
The 2025 Watch List from Canada's Drug Agency (NCBI Bookshelf, NBK613808) identifies AI technologies with the greatest near-term potential to reshape healthcare, noting that 'AI technologies have the potential to significantly transform health care systems' by increasing efficiency, improving patient outcomes, and enhancing patient experience. The Watch List simultaneously flags legal, ethical, environmental, and social implications requiring proactive governance.
2.4 AI in Surgery
Surgical AI encompasses robotic assistance, intraoperative image analysis, tissue recognition, and post-operative outcome prediction. While AI-assisted surgical robotics remains predominantly in specialist academic centres in high-income countries, evidence is accumulating on its safety and efficacy. Machine learning models trained on large surgical video datasets can now perform real-time instrument tracking, anatomical landmark recognition, and procedural step identification, supporting both intraoperative guidance and surgical training.
3. ARTIFICIAL INTELLIGENCE IN DRUG DISCOVERY AND PRECISION MEDICINE
3.1 Revolutionising the Drug Development Pipeline
The traditional pharmaceutical drug development pipeline is notoriously inefficient: it takes an average of 10 to 15 years and over USD 2.5 billion to bring a single new drug to market, with failure rates exceeding 90% in clinical trials. AI is beginning to fundamentally disrupt this paradigm by compressing timelines, reducing costs, and improving the probability of clinical success.
A 2025 systematic review published in PMC (PMC12406033) synthesised peer-reviewed studies published between 2014 and 2024 and found that 'the use of AI lowers costs, shortens the time required for drug development, and enhances the predictive capability.' AI technologies are particularly impactful in three stages:
Target identification and validation: AI analyses genomic, proteomic, and transcriptomic datasets to identify novel disease targets with greater speed and precision than conventional bioinformatics.
Molecular design and screening: Generative AI models propose novel molecular structures optimised for potency, selectivity, and pharmacokinetic properties, dramatically reducing the number of compounds that must be synthesised and tested.
Clinical trial design and patient recruitment: AI algorithms match patients to appropriate clinical trials with high accuracy. A 2025 study of the TrialMatchAI system found it identified matching oncology trials in the top 20 recommendations for 92% of patients, with over 90% accuracy in eligibility classification.
3.2 Landmark AI-Discovered Drug Candidates
Concrete examples of AI's impact on drug discovery are now entering the clinical literature. Insilico Medicine identified a novel target for idiopathic pulmonary fibrosis and advanced a drug candidate (ISM001-055) from target identification through preclinical trials in just 18 months at a cost of approximately USD 150,000 — a process that conventionally requires four to six years and hundreds of millions of dollars (PMC12298131). Positive Phase IIa results for this compound were reported in 2024/2025 (Nature Medicine, 2025; ScienceDirect, 2025).
Exscientia, in partnership with Sumitomo Dainippon Pharma, developed a novel small-molecule drug candidate (DSP-1181) for obsessive-compulsive disorder (OCD) in under 12 months — the first AI-designed molecule to enter human clinical trials. The 2025 merger of Recursion Pharmaceuticals and Exscientia integrated phenomic screening with automated precision chemistry into a full end-to-end discovery platform.
Regulatory progress is also advancing. In December 2025, the FDA qualified its first AI-based tool approved for use in drug development clinical trials: a cloud-based platform assisting pathologists in scoring liver biopsies in NASH/MASH trials — representing formal regulatory acceptance of AI in the drug development process (Drug Target Review, 2026). Earlier, in January 2025, the FDA issued draft guidance titled Considerations for the Use of Artificial Intelligence to Support Regulatory Decision-Making for Drug and Biological Products, establishing a seven-step credibility assessment framework.
3.3 AlphaFold and Protein Structure Prediction
A watershed moment in AI-driven life sciences was the publication of AlphaFold by DeepMind (Nature, 2021) and the subsequent release of AlphaFold 3 (Nature, 2024), which accurately predicted the three-dimensional structures of proteins and their interactions with nucleic acids, small molecules, and other proteins. This capability — solving a problem that had eluded structural biology for 50 years — has opened entirely new pathways for rational drug design by enabling researchers to visualise how candidate molecules interact with their targets at atomic resolution.
4. ARTIFICIAL INTELLIGENCE IN PUBLIC HEALTH
4.1 Disease Surveillance and Early Warning Systems
One of the most critical — and demonstrably life-saving — applications of AI in public health is the enhancement of disease surveillance and outbreak early warning. Traditional surveillance systems rely on passive reporting through healthcare facilities, creating inherent delays between the onset of community transmission and official detection. AI overcomes this limitation by enabling continuous, multi-source, real-time analysis of heterogeneous data streams.
A 2025 systematic review published in Frontiers in Public Health (Mendes et al., 2025, PMC12343694) conducted a comprehensive literature search across PubMed/MEDLINE, Science Direct, IEEE Xplore, and Google Scholar covering January 2017 to December 2024. The review concluded that 'AI has a transformative potential to revolutionise public health by addressing critical challenges in disease prevention, outbreak detection, and countermeasures distribution.' Current efforts are being directed toward integrating heterogeneous data sources — including EHRs, social media, news feeds, genomic databases, climate data, and syndromic surveillance — to detect anomalies indicative of emerging outbreaks.
A related WHO-affiliated study published in BMC Infectious Diseases (Tornimbene et al., 2025, PMC11887143) — produced by researchers at the WHO Hub for Pandemic and Epidemic Intelligence in Berlin — examined AI's role in disease surveillance through the lens of key global health organisations, including Boston Children's Hospital and Médecins Sans Frontières (MSF). The study affirmed that AI's 'utility in clinical care, particularly in diagnostics, medication discovery, and data processing, has resulted in improvements that may also benefit public health surveillance.' The COVID-19 pandemic was identified as a major accelerant of AI-driven surveillance development.
AI-powered early warning platforms have been operationally validated. BlueDot, a Canadian AI epidemiology platform, identified unusual pneumonia clusters in Wuhan, China, and issued alerts to its clients on 31 December 2019 — days before the WHO's official public notification of what would become COVID-19. Machine learning models have been trained to predict malaria with 90% accuracy using support vector machines, random forests, and CatBoost on CDC parasite infection reports and PubMed case abstracts, identifying nationality and travel destination as key predictors (PMC12711821).
4.2 Pandemic Preparedness and Response
The COVID-19 pandemic revealed both the extraordinary utility of AI in emergency public health and the magnitude of the preparedness gaps that remain. AI tools contributed meaningfully across the pandemic response spectrum: contact tracing applications, AI-driven diagnostics from chest CT scans, NLP-powered literature mining to accelerate research synthesis, molecular modelling to identify potential therapeutics, and epidemiological models to inform non-pharmaceutical interventions.
A 2024 scoping review published in SAGE Journals (El Morr et al., 2024) examined 33 peer-reviewed studies on AI-based epidemic and pandemic early warning systems, finding that AI can 'identify patterns in data that signal the onset of epidemics and pandemics.' The review identified key implementation challenges including data quality, standardisation, ethical governance, and the need for multidisciplinary collaboration between AI developers, epidemiologists, and public health authorities.
WHO's 2020 Global Strategy on Digital Health 2020–2025, adopted by the 73rd World Health Assembly, established a strategic vision for AI and digital health as central pillars of pandemic preparedness and the achievement of Sustainable Development Goal (SDG) 3 — ensuring healthy lives for all. The strategy identified four objectives: global collaboration and knowledge transfer; digital health infrastructure; country capacity strengthening; and governance and ethics.
4.3 Maternal and Child Health
Maternal and child mortality remain among the most persistent inequities in global health. According to UNICEF's 2024 Levels and Trends in Child Mortality Report, 'the global effort to reduce child mortality has yielded extraordinary results over the last 30 years,' yet progress 'is at risk of stagnation or even reversal due to a convergence of growing threats: crises, conflict, economic instability, fragile health systems and shrinking donor funding.'
UNICEF data on maternal mortality (2024) reveals that the lifetime risk of maternal death in low-income countries was 1 in 66 in 2023, compared to approximately 1 in 8,000 in high-income countries. Women in sub-Saharan Africa face the highest lifetime risk — 1 in 55 — approximately 250 times higher than in Western Europe (1 in 14,000). WHO has reported that 95% of maternal deaths in developing countries could be prevented with access to appropriate care (Premier Science, 2026, citing WHO Maternal Mortality Fact Sheet 2024).
AI is emerging as a promising tool to address these disparities. A 2025 review published in Frontiers in Global Women's Health (Hoodbhoy et al., 2025, PMC12436304) examined AI for improving maternal and neonatal health in low-resource settings, finding emerging applications in antenatal risk stratification, fetal monitoring, ultrasound interpretation, and postnatal complication prediction. AI-powered ultrasound tools can be operated by community health workers with limited specialist training, extending diagnostic reach into rural and remote areas.
UNICEF's Venture Fund has supported AI start-ups developing solutions for child health, including vaccine coverage prediction, malnutrition screening, and early childhood development monitoring. UNICEF's 2024 Digital Health and Information Systems Annual Report confirms that digital health tools are contributing to 'improving service delivery, enhancing data-driven decision-making, augmenting human resource capacities and strengthening health programs, including those in fragile settings.'
4.4 Non-Communicable Diseases and Mental Health
Non-communicable diseases (NCDs) — including cardiovascular disease, cancer, diabetes, and chronic respiratory disease — account for approximately 74% of all global deaths annually (WHO). Fifteen million premature NCD deaths occur each year, with 85% occurring in LMICs (Springer Nature, 2025). AI is being applied to NCD prevention, early detection, and management through predictive risk modelling using EHR data, wearable device integration for continuous monitoring, and personalised treatment recommendation systems. In mental health, NLP-powered tools are being developed to detect depression, anxiety, and suicidality from clinical notes and patient-reported data, expanding access to mental health screening in resource-constrained settings.
5. ETHICAL FRAMEWORK: WHO'S SIX PRINCIPLES FOR AI IN HEALTH
The ethical governance of AI in health is not an ancillary concern — it is foundational to whether AI achieves beneficial outcomes or exacerbates harm. In June 2021, WHO published its landmark report Ethics and Governance of Artificial Intelligence for Health, described as 'the first consensus report on AI ethics in healthcare settings' (WHO, 2021). The report was developed over two years of consultation by a panel of 20 leading international experts in ethics, digital technology, law, and human rights.
The WHO report identified six core ethical principles, described as 'the first specifically geared toward AI in health with international scope' (PMC11426405), which should guide all actors in the health AI ecosystem — governments, developers, regulators, providers, and patients:
Principle 1: Protect Autonomy
In healthcare contexts, this requires that humans remain in control of healthcare systems and medical decisions. Patient privacy and confidentiality must be protected, and individuals must provide valid informed consent through appropriate legal frameworks for data protection. AI should support, not supplant, human clinical judgment.
Principle 2: Promote Human Well-Being, Safety, and the Public Interest
AI systems must not harm people. Developers must satisfy regulatory requirements for safety, accuracy, and efficacy within well-defined clinical use cases. Quality control and continuous quality improvement mechanisms must be embedded into AI systems throughout their operational life. AI must not result in mental or physical harm that could be avoided through alternative approaches.
Principle 3: Ensure Transparency, Explainability, and Intelligibility
AI systems for health must be transparent in their operation and understandable to the clinicians and patients who use or are affected by them. The rationale behind AI-generated recommendations must be communicable in terms meaningful to clinical decision-making. 'Black box' algorithms that cannot be interrogated or explained are fundamentally incompatible with responsible clinical deployment.
Principle 4: Foster Responsibility and Accountability
AI does not eliminate human responsibility. The WHO guidance clarifies that while AI can assist in clinical decision-making, 'the ultimate responsibility for clinical outcomes remains with human practitioners and the institutions that deploy the technology.' Clear legal and ethical accountability frameworks must be established, including liability regimes ensuring that individuals harmed by AI have access to appropriate redress.
Principle 5: Ensure Inclusiveness and Equity
AI for health must be designed to encourage the widest possible equitable use and access, irrespective of age, sex, gender, income, race, ethnicity, sexual orientation, ability, or other characteristics protected under human rights frameworks. Inclusiveness requires proactive efforts to prevent AI from widening existing health inequities between and within populations.
Principle 6: Promote AI That Is Responsive and Sustainable
AI must be designed and deployed in ways that are responsive to the evolving needs of health systems and populations, and sustainable in terms of economic, environmental, and social impact. Governments and companies must address workforce disruptions, including training health workers to adapt to AI and managing potential job displacement from automated systems.
6. CHALLENGES AND CRITICAL CONCERNS
6.1 Algorithmic Bias and Health Equity
Algorithmic bias represents one of the most serious and under-investigated threats to the ethical deployment of AI in public health. A 2025 analysis published in Frontiers in Public Health (Joseph, 2025, PMC12325396) characterises algorithmic bias in public health AI as 'a silent threat to equity in low-resource settings,' noting that 'AI systems are only as effective as the data used to train them and the assumptions under which they are created.'
The foundational problem is representational: most AI training datasets are drawn disproportionately from populations in high-income countries. A 2024 review of 91 clinical text datasets found that 73% of the data originated from the Americas and Europe — regions representing only 22% of the global population — with more than half of all datasets in English (PMC12040885). This data inequity means that AI systems trained on such datasets will systematically underdiagnose, misclassify, or ignore patterns in non-conforming populations, amplifying rather than reducing health disparities.
The consequences for LMICs are particularly severe. A 2025 Springer Nature chapter on AI in LMIC health systems identifies a projected shortfall of ten million health-care workers by 2030 and 15 million premature NCD deaths per year, 85% of which occur in LMICs, yet notes that AI tools developed in high-income countries routinely fail to perform adequately when deployed without adaptation in these settings. Context-specific data strategies, ethical governance, and health-system integration are identified as essential preconditions for beneficial AI deployment in LMICs.
6.2 Data Privacy, Security, and Governance
Health data is among the most sensitive categories of personal information. The training and operation of AI systems in healthcare require access to large, richly detailed datasets, creating inherent tensions with individual privacy rights and national data sovereignty. These tensions are compounded in cross-border AI development, where data from health systems in LMICs may be used to train algorithms that are then sold back to those same countries as commercial products.
The WHO's 2021 ethics guidance, the European Union's AI Act (2024) — the world's first comprehensive legal framework for AI — and the WHO's 2024 guidance on ethics and governance of large multi-modal models all establish requirements for responsible data governance. The 2024–2025 WHO European Region survey on AI for healthcare found significant variation in the maturity of national legal and ethical frameworks, data governance models, and workforce readiness across Member States.
6.3 Regulatory Uncertainty
The regulatory landscape for AI medical devices and AI-assisted clinical tools is evolving rapidly but remains immature in most jurisdictions. As of 2025, the FDA has cleared AI algorithms primarily in radiology and cardiology for specific detection tasks. The FDA's January 2025 draft guidance on AI for regulatory decision-making represents an important step, but the 'dynamic nature of AI (with potential updates or drifts) does not fit neatly into current regulatory frameworks that assume static devices' (PMC12468291). Post-market surveillance requirements, adaptive approval pathways, and international regulatory harmonisation are active areas of policy development.
6.4 Health Workforce Readiness and the Digital Divide
The benefits of AI cannot be realised in health systems that lack the digital infrastructure, connectivity, electricity supply, and digitally trained workforce necessary to deploy and use AI tools. WHO's 2024–2025 survey of the European Region found substantial variation in workforce preparedness even among relatively well-resourced countries. In LMICs, challenges are compounded: limited internet penetration, absence of digital health records, chronic underfunding of health systems, and inadequate training for health workers create barriers that no AI algorithm alone can overcome.
UNICEF's 2024 Digital Health and Information Systems Annual Report acknowledges that 'challenges such as funding constraints, the need for greater capacity, and improving global coordination remain,' even as digital solutions are expanded into primary healthcare and fragile settings.
6.5 Risk of Over-Reliance and Loss of Clinical Judgment
A nuanced but serious concern is the risk that AI tools — particularly those with high apparent accuracy — may erode clinical judgment and professional skills over time, creating systems in which clinicians defer to algorithmic outputs without adequate critical appraisal. The WHO's ethical guidance and the broader academic literature consistently emphasise that AI should 'complement, not replace, doctors and healthcare providers' (Nature Digital Medicine, 2023, cited in npj Digital Medicine, 2025). Maintaining robust clinical training, fostering AI literacy among healthcare professionals, and embedding human oversight into AI-assisted workflows are essential safeguards.
7. OPPORTUNITIES AND THE PATH FORWARD
7.1 Achieving Universal Health Coverage
AI holds genuine promise as a tool for advancing Universal Health Coverage (UHC) — the SDG commitment that all people should have access to quality health services without financial hardship. In regions where physician-to-population ratios are critically low, AI-powered diagnostic tools, clinical decision support, and telemedicine platforms can extend the effective reach of limited specialist capacity. Community health workers equipped with AI-assisted mobile diagnostic tools can provide screening and triage functions that would otherwise be unavailable.
7.2 Global Collaboration and Knowledge Transfer
WHO's Global Strategy on Digital Health 2020–2025 explicitly identifies global collaboration and the transfer of knowledge on digital health as a strategic priority. The Ministry of Electronics and Information Technology of India and WHO jointly launched a global call for AI in health abstracts in October 2025, seeking impactful and scalable applications of AI in health systems for a forthcoming Casebook on AI health use cases across the Global South — recognising that innovations developed and validated in low-resource settings may be more relevant and transferable than those developed in high-income country contexts (NewsonAir, 2025).
WHO's Global Initiative on AI for Health (GI-AI4H) — co-founded with the International Telecommunication Union (ITU) and the World Intellectual Property Organization (WIPO) — is advancing standardised guidelines, knowledge-sharing frameworks, and pilot initiatives across member states to ensure that AI in health develops through coordinated global rather than fragmented national efforts.
7.3 Inclusive and Participatory AI Development
A 2025 paper in PMC (PMC12766452) argues for 'redistributing ethical authority to affected communities, particularly in low- and middle-income countries, potentially through participatory councils with decision power over evaluation metrics, equity constraints, and deployment.' Centring the lived experience of the communities that AI health tools are designed to serve — through co-design, community engagement, and participatory monitoring — can simultaneously strengthen technical performance, contextual validity, ethical accountability, and social trust.
7.4 Investment in Data Infrastructure and Digital Literacy
Realising AI's potential in global health requires sustained, coordinated investment in foundational infrastructure: digital health record systems, interoperable data standards, reliable connectivity, and educational programmes that build AI literacy among health professionals, policymakers, and communities. These investments must be accompanied by equity-centred data governance frameworks that ensure populations in LMICs are represented in training data and retain meaningful sovereignty over their health information.
8. CONCLUSION
Artificial intelligence is not a future prospect for medicine and public health — it is an active and accelerating present reality. It is improving diagnostic accuracy in radiology and pathology, compressing drug discovery timelines that have historically spanned decades, enabling real-time disease surveillance that can detect outbreaks before traditional systems would register them, and extending diagnostic and clinical decision support capabilities into settings previously bereft of specialist expertise.
Yet the evidence reviewed in this article makes clear that the trajectory of AI in health is neither uniformly beneficial nor inevitable. The same technology that can detect cancer earlier in a well-resourced hospital can perpetuate diagnostic inequity if trained on unrepresentative data. The same algorithmic efficiency that accelerates drug discovery in a high-income laboratory may never reach the populations who bear the greatest burden of disease. The WHO's six ethical principles — protecting autonomy, promoting safety and well-being, ensuring transparency, fostering accountability, ensuring equity, and promoting sustainability — are not aspirational abstractions. They are essential operating requirements for AI that serves humanity rather than a subset of it.
The global health community stands at a pivotal moment. The policy frameworks, regulatory architectures, investment decisions, and governance structures being established now will shape whether AI in health delivers on its extraordinary promise for all people — or widens the chasms of inequity that existing health systems have failed to close. The WHO, UNICEF, national governments, academic institutions, technology developers, and civil society must act with urgency, rigour, and a shared commitment to health as a universal human right.
REFERENCES
1. World Health Organization (WHO). (2021). Ethics and Governance of Artificial Intelligence for Health. Geneva: WHO. https://www.who.int/publications/i/item/9789240029200
2. World Health Organization (WHO). (2021, June 28). WHO issues first global report on AI in health and six guiding principles for its design and use. WHO News Release. https://www.who.int/news/item/28-06-2021-who-issues-first-global-report-on-ai-in-health-and-six-guiding-principles-for-its-design-and-use
3. World Health Organization (WHO). (2024). Ethics and Governance of Artificial Intelligence for Health: Guidance on Large Multi-Modal Models. Geneva: WHO. https://iris.who.int/
4. World Health Organization (WHO). (2025). Artificial Intelligence is Reshaping Health Systems: State of Readiness across the WHO European Region. Based on findings from the 2024–2025 Survey on AI for Health Care. Copenhagen: WHO Regional Office for Europe. https://www.who.int/europe/publications/i/item/WHO-EURO-2025-12707-52481-81028
5. World Health Organization (WHO). (2020). Global Strategy on Digital Health 2020–2025. Geneva: WHO. https://apps.who.int/iris/handle/10665/344249
6. World Health Organization (WHO). (2023, October 21). WHO Outlines Considerations for Regulation of Artificial Intelligence for Health. WHO News Release. https://www.who.int/europe/news/item/21-10-2023-who-outlines-considerations-for-regulation-of-artificial-intelligence-for-health
7. UNICEF. (2024). Levels and Trends in Child Mortality 2024. New York: UNICEF. https://data.unicef.org/resources/levels-and-trends-in-child-mortality-2024/
8. UNICEF. (2024). Maternal Mortality Rates and Statistics. UNICEF Data. https://data.unicef.org/topic/maternal-health/maternal-mortality/
9. UNICEF. (2024). Digital Health and Information Systems: 2024 Annual Report. UNICEF Digital Impact. https://www.unicef.org/digitalimpact/reports/unicef-digital-health-and-information-systems-2024-annual-report
10. UNICEF Office of Innovation. Four Startups Harnessing Artificial Intelligence to Strengthen Healthcare Systems for Children. https://www.unicef.org/innovation/stories/four-startups-harnessing-artificial-intelligence-strengthen-healthcare-systems-children
11. Strubchevska, O., Kozyk, M., Kozyk, A., & Strubchevska, K. (2024). The Role of Artificial Intelligence in Diagnostic Radiology. Cureus, 16(10), e72173. https://doi.org/10.7759/cureus.72173. PMC11582495.
12. Bhandari, A. (2024). Revolutionizing Radiology with Artificial Intelligence. Cureus, 16(10), e72646. https://doi.org/10.7759/cureus.72646. PMC11521355.
13. Banerjee, A., Shan, H., & Feng, R. (2025). Artificial Intelligence Applications for Cancer Diagnosis in Radiology. Frontiers in Radiology. https://doi.org/10.3389/fradi.2025.1493783. PMC11813860.
14. Singh, P., et al. (2024). Artificial Intelligence-Assisted Medical Imaging Solutions for Integrating Pathology and Radiology Automated Systems. Frontiers in Medicine. https://doi.org/10.3389/fmed.2024.1447294. PMC11236738.
15. [Authors]. (2025). Artificial Intelligence in Clinical Medicine: Challenges Across Diagnostic Imaging, Clinical Decision Support, Surgery, Pathology, and Drug Discovery. PMC12468291.
16. Canada's Drug Agency. (2025). 2025 Watch List: Artificial Intelligence in Health Care. NCBI Bookshelf. NBK613808.
17. Mendes, V.I.S., et al. (2025). Harnessing Artificial Intelligence for Enhanced Public Health Surveillance: A Narrative Review. Frontiers in Public Health. https://doi.org/10.3389/fpubh.2025.1601151. PMC12343694.
18. Tornimbene, B., et al. (2025). Harnessing the Power of Artificial Intelligence for Disease-Surveillance Purposes. BMC Infectious Diseases. https://doi.org/10.1186/s12919-025-00320-w. PMC11887143.
19. Villanueva-Miranda, I., Xiao, G., & Xie, Y. (2025). Artificial Intelligence in Early Warning Systems for Infectious Disease Surveillance: A Systematic Review. Frontiers in Public Health. https://doi.org/10.3389/fpubh.2025.1609615. PMC12230060.
20. El Morr, C., et al. (2024). AI-Based Epidemic and Pandemic Early Warning Systems: A Systematic Scoping Review. Health Informatics Journal, 30. https://doi.org/10.1177/14604582241275844.
21. Anjaria, P., et al. (2023). Artificial Intelligence in Public Health: Revolutionising Epidemiological Surveillance for Pandemic Preparedness and Equitable Vaccine Access. Vaccines, 11(7), 1154. https://doi.org/10.3390/vaccines11071154. PMC10383160.
22. [Authors]. (2025). Artificial Intelligence in Epidemic Watch: Revolutionizing Infectious Diseases Surveillance. PMC12711821.
23. Hoodbhoy, Z., et al. (2025). Use of Artificial Intelligence to Improve Maternal and Neonatal Health in Low-Resource Settings. Frontiers in Global Women's Health. https://doi.org/10.3389/fgwh.2025.1675578. PMC12436304.
24. [Authors]. (2025). Can AI Bridge or Widen Maternal Health Inequities? PMC12445195.
25. Premier Science. (2026). The Role of Artificial Intelligence in Enhancing Maternal and Child Health Through Digital Health Initiatives in Resource-Limited Settings: A Narrative Review. https://premierscience.com/pjai-25-1238/
26. [Authors]. (2025). Transformative Role of Artificial Intelligence in Drug Discovery and Translational Medicine. PMC12406033.
27. [Authors]. (2025). From Lab to Clinic: How Artificial Intelligence Is Reshaping Drug Discovery Timelines and Industry Outcomes. PMC12298131.
28. [Authors]. (2025). Leading Artificial Intelligence-Driven Drug Discovery Platforms: 2025 Landscape and Global Outlook. ScienceDirect.
29. Drug Target Review. (2026). AI in Drug Discovery: 2025 in Review. https://www.drugtargetreview.com/article/192951/ai-in-drug-discovery-2025-in-review/
30. Abramson, J., et al. (2024). Accurate Structure Prediction of Biomolecular Interactions with AlphaFold 3. Nature, 630, 493–500.
31. Jumper, J., et al. (2021). Highly Accurate Protein Structure Prediction with AlphaFold. Nature, 596, 583–589.
32. Joseph, J. (2025). Algorithmic Bias in Public Health AI: A Silent Threat to Equity in Low-Resource Settings. Frontiers in Public Health. https://doi.org/10.3389/fpubh.2025.1643180. PMC12325396.
33. [Authors]. (2025). The Imperative of Diversity and Equity for the Adoption of Responsible AI in Healthcare. PMC12040885.
34. [Authors]. (2025). Rethinking Expertise in Artificial Intelligence Ethics for Global Health. PMC12766452.
35. López, D.M., et al. (2022). Challenges and Solutions for Transforming Health Ecosystems in Low- and Middle-Income Countries through Artificial Intelligence. Frontiers in Medicine. https://doi.org/10.3389/fmed.2022.958097. PMC9755337.
36. [Authors]. (2025). Artificial Intelligence in Healthcare Systems of Low- and Middle-Income Countries: Opportunities, Challenges, and Pathways to Ethical Integration. Springer Nature.
37. [Authors]. (2025). Transforming Healthcare through Just, Equitable and Quality Driven Artificial Intelligence Solutions in South Asia. npj Digital Medicine, 8, 139. https://doi.org/10.1038/s41746-025-01534-0.
38. Goodman, K.W., & Litewka, S.G. (2023). Global Health and Big Data: The WHO's Artificial Intelligence Guidance. South African Journal of Science, 119(5-6). PMC11426405.
39. International Agency for Research on Cancer (IARC)/WHO. (2024). Global Cancer Burden Data 2023 / Annual Cancer Report 2024. Lyon: IARC.
40. [Authors]. (2025). Precision Oncology in the Age of AI: Lessons from AI-Driven Drug Discovery and Clinical Translation. BJC Reports. Nature.
41. NewsonAir/Ministry of Electronics and Information Technology & WHO. (2025). MEIT and WHO Launch Global Call for AI in Health Abstracts. https://www.newsonair.gov.in/meit-and-who-launch-global-call-for-ai-in-health-abstracts
42. WHO AI for Good Summit. (2025). Enabling AI for Health Innovation and Access. Geneva: WHO. https://www.who.int/news-room/events/detail/2025/07/11/default-calendar/ai-for-good-summit-2025---enabling-ai-for-health-innovation-and-access
43. WHO Europe. (2026). Artificial Intelligence is Reshaping Health Systems: State of Readiness within EU Health Systems. https://www.who.int/europe/publications/i/item/WHO-EURO-2026-12707-52481-81471
44. US Food and Drug Administration (FDA). (2025, January 6). Considerations for the Use of Artificial Intelligence to Support Regulatory Decision-Making for Drug and Biological Products. Draft Guidance. Silver Spring: FDA.