AI in Global Health and Equity

Global health AI should be evaluated as part of a public health program or service, not as a model transferred between countries. A system that performs well in one setting can fail when the population, data, language, workflow, infrastructure, referral capacity, governance, or maintenance environment differs. The practical question is whether the complete workflow improves a defined public health decision under the conditions where it will operate.

Learning Objectives

This chapter equips readers to:

  • Define the intended decision, current workflow, affected population, and accountable institution before selecting a technology
  • Assess connectivity, electricity, computing, data, interoperability, workforce, and institutional constraints at the intended sites
  • Distinguish model performance from local validity, service impact, equity, and sustainability
  • Compare an AI system with the current workflow, simpler digital tools, and non-AI alternatives
  • Evaluate whether data governance, benefit sharing, capacity building, and exit terms preserve meaningful local authority
  • Assign ownership for validation, procurement, maintenance, monitoring, incident response, fallback, and suspension

Core rule: Evaluate the whole service under local operating conditions. A model is useful only when the required inputs are available, its outputs remain valid, people can act on those outputs, and the service can be governed and maintained.

Decision sequence:

  1. Define the need. Name the decision, current workflow, intended users, affected population, available action, and non-AI baseline.
  2. Measure the setting. Document site-level connectivity, power, hardware, data, language, staffing, referral, financing, policy, and maintenance constraints.
  3. Test transferability. Review evidence from relevant populations and settings, then require local validation and workflow testing proportionate to risk.
  4. Compare complete options. Assess benefits, harms, acceptability, feasibility, resource use, and equity across the AI proposal and viable alternatives.
  5. Assign authority and lifecycle duties. Establish who controls data and deployment, who benefits, who maintains and monitors the system, and who can suspend or end its use.

Critical takeaways:

  • Country income group is not a deployment specification. Conditions vary within countries, districts, facilities, and communities.
  • The World Health Organization states that digital interventions are not substitutes for functioning health systems.
  • External benchmark performance does not establish local effectiveness, fairness, or service impact.
  • A simpler system can be the stronger option when it is more reliable, maintainable, affordable, and compatible with the current workflow.
  • Local institutions and affected communities need meaningful authority over data, implementation, monitoring, benefits, and exit conditions.

Introduction: Evaluate the Service, Not Only the Model

Evidence for an algorithm and evidence for a public health intervention are not interchangeable. Performance measured in one population, device environment, language, or workflow does not establish that the same system will improve decisions elsewhere. Evaluation must include the people, infrastructure, information systems, operating procedures, downstream services, and governance needed to convert an output into action.

WHO’s guideline on digital interventions for health system strengthening recommends examining benefits, harms, acceptability, feasibility, resource use, and equity. Its AI ethics and governance guidance identifies principles covering autonomy; well-being, safety, and the public interest; transparency; accountability; inclusiveness and equity; and responsiveness and sustainability. Together, these sources support a health-system assessment rather than technology transfer based on model accuracy alone.

Constraints differ within as well as between countries. A national connectivity estimate does not show whether a particular facility has reliable power, affordable data, compatible records, trained staff, referral capacity, or authority to maintain a system. Site-level measurement and local participation are therefore prerequisites for an implementation decision.

This chapter provides the overview and decision path. Detailed methods remain in Context-Appropriate Global Health AI, Global Health AI Equity and Local Capacity, Global Health AI Data Governance, and Global Health AI Exercises.

The Digital Divide in Global Health AI

The digital divide is not a single connectivity measure. It includes the infrastructure, information systems, workforce, financing, governance, and service capacity required to use an AI output safely and consistently. Weakness in one layer can make the complete workflow unavailable even when the model itself runs.

Interdependent Infrastructure Constraints

1. Connectivity Divide

Connectivity should be assessed at the point of service, including uptime, bandwidth, latency, cost, device access, and the ability to continue safely during an outage.

Key findings: In 2025, ITU estimated that 2.2 billion people remained offline. Internet use was 85% in urban areas versus 58% in rural areas, and 77% among men versus 71% among women (ITU, 2025).

2. Computing Infrastructure Divide

Most AI models assume access to: - Cloud computing (AWS, Azure, GCP) - requires internet + payment infrastructure - GPUs for training ($5,000-$50,000 per unit) - prohibitive for many institutions - Continuous electricity - The World Bank reported in 2026 that 655 million people lacked electricity access (World Bank, 2026)

Example: Training a medical imaging model

High-income setting:

# Standard GPU training approach
import tensorflow as tf

# Assumes: V100 GPU ($10,000), 24/7 electricity, high-speed internet
model = tf.keras.applications.ResNet50(weights='imagenet', include_top=False)

# Train on 100,000 chest X-rays
history = model.fit(
 train_dataset, # Loaded from cloud storage
 epochs=100,
 batch_size=32, # Requires 16GB+ GPU memory
 validation_data=val_dataset
)

LMIC reality: - No GPU available → Training time: 100 hours (GPU) → 2,000+ hours (CPU) - Intermittent electricity → Training interrupted, progress lost - Limited internet → Cannot download large pre-trained models or datasets - Cost: GPU electricity alone might exceed annual healthcare budget for facility

3. Data Infrastructure Divide

# Comparison of health data infrastructure

data_infrastructure = {
 'Feature': [
  'Electronic Health Records',
  'Health Information Exchange',
  'Unique Patient Identifiers',
  'Structured Data',
  'Interoperability Standards',
  'Data Quality Monitoring',
  'Real-time Surveillance Systems'
 ],
 'High-Income (%)': [95, 80, 90, 75, 70, 85, 90],
 'Upper-Middle (%)': [45, 25, 40, 35, 30, 35, 50],
 'Lower-Middle (%)': [15, 8, 15, 12, 10, 15, 25],
 'Low-Income (%)': [5, 2, 5, 5, 3, 5, 10]
}

df = pd.DataFrame(data_infrastructure)

# Visualize infrastructure gap
fig = px.bar(df, x='Feature',
    y=['High-Income (%)', 'Upper-Middle (%)',
    'Lower-Middle (%)', 'Low-Income (%)'],
    title='Health Data Infrastructure by Country Income Level',
    barmode='group',
    height=500)

fig.update_layout(xaxis_tickangle=-45)
fig.show()

Implications for AI: - No EHRs → No training data for predictive models - Paper records → Labor-intensive data entry, high error rates - No patient identifiers → Cannot track patients across visits - Unstructured data → Requires complex NLP preprocessing - No interoperability → Data silos prevent full analysis - Weak surveillance infrastructure → Event-based surveillance systems (HealthMap, WHO EIOS) detected only 9.2% of influenza outbreaks within a two-week window in a 24-country evaluation, with LMIC performance consistently weaker (Ganser et al., 2022). See AI in Disease Surveillance for details.

4. Human Capacity Divide

Data scientists per capita: - USA: ~50 per 100,000 population - Sub-Saharan Africa: ~0.2 per 100,000 population

Healthcare workers per capita: - WHO recommendation: 44.5 per 10,000 population - Sub-Saharan Africa average: 13.2 per 10,000 - Shortage of 6.4 million healthcare workers in Africa

The Paradox of Need

Settings with the greatest potential AI impact often have the least capacity to develop, deploy, and maintain AI systems. This creates a vicious cycle where:

  1. Limited resources → Limited AI capacity
  2. Limited AI capacity → Dependence on external solutions
  3. External solutions → Not context-appropriate
  4. Poor outcomes → Distrust of technology
  5. Distrust → Reduced investment → Back to #1

Breaking this cycle requires intentional capacity building and locally-led innovation.

However, user sentiment data complicates the assumption that distrust is uniform. Anthropic’s survey of 80,508 AI users across 159 countries found that respondents in Sub-Saharan Africa, Central Asia, South Asia, and Latin America expressed significantly more optimism about AI’s potential than those in North America, Western Europe, and Oceania (Anthropic, 2025). Developing-region respondents emphasized entrepreneurship and education as primary aspirations for AI, while high-income-country respondents focused on life management and personal transformation. This suggests that step 4 in the cycle above (poor outcomes leading to distrust) may be less entrenched than assumed: populations facing the greatest healthcare access gaps are often the most receptive to AI-assisted solutions when tools are designed for their context.

Context-Appropriate AI Design

The detailed material is maintained in Context-Appropriate Global Health AI. This section anchor remains here for continuity.

Algorithmic Fairness Across Populations

The detailed material is maintained in Global Health AI Equity and Local Capacity. This section anchor remains here for continuity.

Data Governance in Global Health

The detailed material is maintained in Global Health AI Data Governance. This section anchor remains here for continuity.

Key Takeaways

  1. AI amplifies inequity by default - Without intentional equity focus, AI widens health disparities by concentrating benefits in privileged populations.

  2. Design for constraints, not ideal conditions:

  • Offline-first - Works without continuous connectivity
  • Low-power - Runs on battery-powered devices
  • Robust to poor data - Handles low-quality images, missing values, limited labels
  • Culturally appropriate - Localized language, terminology, and cultural norms

  1. Success stories demonstrate feasibility - Retinopathy screening (India), TB detection (Africa), malaria diagnosis (SE Asia) show AI can work in resource-limited settings when thoughtfully designed.

  2. Algorithmic fairness requires active work:

  • Diverse training data from multiple populations
  • Fairness-aware training with constraints
  • Regular auditing across demographic groups
  • Local fine-tuning for specific contexts
  1. Build sustainable local capacity:
  • Train local data scientists (awareness → literacy → application → development → research)
  • Invest in local infrastructure
  • Share code, models, and knowledge
  • Plan for long-term local ownership
  1. Data governance prevents exploitation:
  • Data sovereignty - Source retains control
  • Benefit sharing - Co-authorship, IP sharing, revenue sharing
  • Equitable partnerships - Avoid data colonialism
  • Capacity building - Knowledge transfer, not extraction

  1. Context-appropriateness beats sophistication - Simple SMS-based system with local ownership often outperforms sophisticated cloud system requiring continuous connectivity.

  2. Measure what matters:

  • Not just accuracy, but also context-appropriateness
  • Not just performance, but also sustainability
  • Not just deployment, but also equity and local ownership

  1. Infrastructure is foundational - No amount of sophisticated AI can overcome lack of electricity, internet, or digital health systems.

  2. Global health AI should benefit those who need it most - If AI only works for the wealthy and connected, we have failed.

Hands-On Exercise: Assess AI Tool for Your Context

Objective: Evaluate whether an AI health tool is appropriate for a resource-limited setting.

Scenario: You are a public health official in a lower-middle-income country. An international NGO is offering to deploy an AI-powered disease surveillance system in your country. Assess whether this system is appropriate for your context.

Part 1: Context Assessment (15 minutes)

Define your context:

my_context = {
 'country': 'Example Country',
 'income_level': 'Lower-middle',
 'infrastructure': {
  'internet_penetration': 0.45, # 45% population
  'electricity_reliability': 0.70, # Reliable 70% of time
  'mobile_phone_penetration': 0.80, # 80% population
  'ehr_penetration': 0.10, # 10% facilities have EHR
  'health_workers_per_10k': 15 # Below WHO recommendation of 44.5
 },
 'priority_health_issues': ['Malaria', 'TB', 'Maternal mortality', 'Malnutrition'],
 'languages': ['English', 'Local Language 1', 'Local Language 2'],
 'existing_surveillance': 'Paper-based reporting from facilities to district to national level'
}

Questions: 1. What are your main infrastructure constraints? 2. What are your priority health problems? 3. What is your current surveillance system?

Part 2: AI Tool Assessment (20 minutes)

The proposed AI tool:

proposed_tool = {
 'name': 'SmartSurveillance AI',
 'purpose': 'Real-time disease outbreak detection',
 'data_sources': [
  'Electronic health records',
  'Laboratory results',
  'Social media monitoring',
  'Weather data',
  'Google search trends'
 ],
 'requirements': {
  'internet': 'High-speed internet required',
  'hardware': 'Cloud-based (AWS), requires continuous internet',
  'data': 'Requires structured EHR data',
  'training': '2-week training for data analysts',
  'cost': '$50,000 setup + $20,000/year subscription',
  'support': 'Email support (24-48 hour response)'
 },
 'languages': ['English'],
 'development': 'Developed in USA, trained on US data',
 'evidence': 'Tested in USA, UK, Australia'
}

Assess appropriateness:

class ContextAppropriatenessChecker:
 """Check if AI tool is appropriate for context"""

 @staticmethod
 def check_requirements(context, tool):
  """Check if context meets tool requirements"""
  checks = {}

  # Infrastructure checks
  checks['internet'] = {
   'required': 'High-speed',
   'available': f"{context['infrastructure']['internet_penetration']:.0%} penetration",
   'appropriate': context['infrastructure']['internet_penetration'] > 0.80
  }

  checks['electricity'] = {
   'required': 'Continuous',
   'available': f"{context['infrastructure']['electricity_reliability']:.0%} reliable",
   'appropriate': context['infrastructure']['electricity_reliability'] > 0.95
  }

  checks['data_infrastructure'] = {
   'required': 'EHR with structured data',
   'available': f"{context['infrastructure']['ehr_penetration']:.0%} facilities with EHR",
   'appropriate': context['infrastructure']['ehr_penetration'] > 0.70
  }

  # Cost check (as % of health budget per capita)
  # Example: Low-middle income ~$100 per capita health spending
  annual_cost = tool['requirements']['cost'].split('+')[1].replace('$','').replace('/year subscription','').replace(',','')
  checks['cost'] = {
   'required': f"${annual_cost}/year",
   'appropriate': 'Requires cost-benefit analysis',
   'note': 'Compare to alternative surveillance approaches'
  }

  # Language check
  checks['language'] = {
   'required': ', '.join(tool['languages']),
   'available': ', '.join(context['languages']),
   'appropriate': any(lang in tool['languages'] for lang in context['languages'])
  }

  return checks

 @staticmethod
 def recommend_alternatives(context, tool, checks):
  """Recommend modifications or alternatives"""
  recommendations = []

  if not checks['internet']['appropriate']:
   recommendations.append({
    'issue': 'Insufficient internet connectivity',
    'recommendation': 'Request offline-first version',
    'alternative': 'Use SMS-based reporting with simpler ML models'
   })

  if not checks['data_infrastructure']['appropriate']:
   recommendations.append({
    'issue': 'Insufficient EHR penetration',
    'recommendation': 'Start with paper-based data entry',
    'alternative': 'Use syndromic surveillance (symptoms only, no lab data)'
   })

  if not checks['language']['appropriate']:
   recommendations.append({
    'issue': 'Language mismatch',
    'recommendation': 'Request localization to [Local Languages]',
    'alternative': 'Develop local language version with local partner'
   })

  # Always recommend local validation
  recommendations.append({
   'issue': 'Tool not validated in local context',
   'recommendation': 'Require pilot study with local data',
   'alternative': 'Collaborate with local university to validate and adapt'
  })

  return recommendations

# Run assessment
checker = ContextAppropriatenessChecker()
checks = checker.check_requirements(my_context, proposed_tool)

print("CONTEXT-APPROPRIATENESS ASSESSMENT")
print("="*50)
for requirement, result in checks.items():
 status = "[OK]" if result.get('appropriate', False) else "[FAIL]"
 print(f"\n{status} {requirement.upper()}:")
 print(f" Required: {result['required']}")
 print(f" Available: {result['available']}")

# Get recommendations
recommendations = checker.recommend_alternatives(my_context, proposed_tool, checks)
print("\n\nRECOMMENDATIONS")
print("="*50)
for i, rec in enumerate(recommendations, 1):
 print(f"\n{i}. {rec['issue']}")
 print(f" Recommendation: {rec['recommendation']}")
 print(f" Alternative: {rec['alternative']}")

Your assessment: 1. Is this tool appropriate for your context as-is? 2. What modifications would make it more appropriate? 3. What alternatives might be better suited to your context?

Part 3: Equity and Governance (20 minutes)

Assess the partnership:

partnership_terms = {
 'data': {
  'data_storage': 'Cloud (AWS, USA servers)',
  'data_ownership': 'NGO retains ownership',
  'data_use': 'NGO can use for research and product improvement',
  'data_sharing': 'NGO can share with partners'
 },
 'capacity_building': {
  'training': '2-week online training',
  'ongoing_support': 'Email support only',
  'local_capacity': 'No plan for local AI capacity'
 },
 'intellectual_property': {
  'ownership': 'NGO owns all IP',
  'customization': 'NGO controls all customization',
  'code_access': 'No access to source code'
 },
 'costs': {
  'setup': '$50,000 (funded by NGO)',
  'subscription': '$20,000/year (funding for 3 years, then country pays)',
  'support': 'Included in subscription'
 },
 'evaluation': {
  'performance': 'NGO evaluates',
  'data': 'NGO publishes findings',
  'feedback': 'Country can provide feedback'
 }
}

# Assess governance equity
print("\nGOVERNANCE EQUITY ASSESSMENT")
print("="*50)

red_flags = []
green_flags = []

# Data control
if 'retains ownership' in partnership_terms['data']['data_ownership']:
 red_flags.append("[FAIL] Country does not control its own health data")
else:
 green_flags.append("[OK] Country retains control of health data")

# Capacity building
if 'No plan' in partnership_terms['capacity_building']['local_capacity']:
 red_flags.append("[FAIL] No sustainable local capacity building")
else:
 green_flags.append("[OK] Plan for sustainable local capacity")

# IP ownership
if 'NGO owns' in partnership_terms['intellectual_property']['ownership']:
 red_flags.append("[FAIL] Country has no IP rights to system built with its data")
else:
 green_flags.append("[OK] Joint IP ownership")

# Sustainability
if 'country pays' in partnership_terms['costs']['subscription']:
 red_flags.append("[FAIL] Dependency on external system with ongoing costs")
else:
 green_flags.append("[OK] Sustainable funding model")

print("\nRED FLAGS:")
for flag in red_flags:
 print(f" {flag}")

print("\nGREEN FLAGS:")
for flag in green_flags:
 print(f" {flag}")

print("\n\nRECOMMENDED NEGOTIATION POINTS:")
print("1. Data sovereignty: Country must retain ownership and control of health data")
print("2. Local storage: Data stored on local servers, not exported")
print("3. Capacity building: Commit to training local team to run system independently")
print("4. IP sharing: Joint ownership of any adaptations or innovations")
print("5. Open source: Access to source code for local customization")
print("6. Sustainability: Plan for local ownership within 3 years")
print("7. Evaluation: Joint evaluation with local researchers as co-authors")

Your assessment: 1. What are the red flags in this partnership? 2. What would you negotiate to make it more equitable? 3. Would you accept this partnership? Why or why not?

Part 4: Alternative Approach (25 minutes)

Design a more appropriate alternative:

Your task: Design a disease surveillance system that: - Works within your infrastructure constraints - Builds local capacity - Is sustainable - Maintains data sovereignty

Example approach:

alternative_system = {
 'name': 'LocalSurveillance',
 'design_principles': [
  'Offline-first',
  'Works with existing paper systems',
  'Low-cost',
  'Locally developed and owned',
  'Builds local capacity'
 ],
 'architecture': {
  'data_collection': 'SMS + simple mobile app (works offline)',
  'data_storage': 'Local servers (Ministry of Health)',
  'analysis': 'Simple ML models (run on laptop/local server)',
  'alerts': 'SMS alerts to health officials',
  'dashboard': 'Web dashboard (works on low bandwidth)'
 },
 'implementation': {
  'phase_1': '6 months - SMS pilot in 2 districts',
  'phase_2': '6 months - Expand to 10 districts, add simple ML',
  'phase_3': '12 months - National rollout'
 },
 'capacity_building': {
  'train_local_developers': '3-month intensive training',
  'train_health_workers': '1-day training on data entry',
  'establish_local_team': 'Hire 5 local data scientists',
  'partner_with_university': 'Joint project with local university'
 },
 'costs': {
  'setup': '$20,000 (servers, initial development)',
  'annual': '$10,000 (maintenance, hosting)',
  'capacity_building': '$50,000 (one-time training investment)',
  'total_5_year': '$120,000 vs $150,000 for proprietary system',
  'after_5_years': 'Fully locally owned and operated'
 },
 'governance': {
  'data_ownership': 'Ministry of Health',
  'code_ownership': 'Open source',
  'sustainability': 'Locally maintained',
  'scalability': 'Can be adapted for other countries'
 }
}

print("ALTERNATIVE SYSTEM: LocalSurveillance")
print("="*50)
print(f"\nDesign Principles:")
for principle in alternative_system['design_principles']:
 print(f" * {principle}")

print(f"\nCost Comparison (5 years):")
print(f" Proprietary system: $150,000")
print(f" LocalSurveillance: $120,000")
print(f" Savings: $30,000")
print(f"\n Plus: Builds local capacity, local ownership, sustainable")

print(f"\nCapacity Building:")
for activity, details in alternative_system['capacity_building'].items():
 print(f" • {activity}: {details}")

Your turn: 1. Design an appropriate AI system for your context 2. How would you implement it? 3. How would you build local capacity? 4. How much would it cost? 5. How is it better than the proprietary alternative?

Check Your Understanding

The detailed material is maintained in Global Health AI Exercises. This section anchor remains here for continuity.

Why can a health AI system fail when transferred to a lower-resource setting?

The target population, disease prevalence, language, clinical workflow, equipment, data definitions, referral capacity, connectivity, power, staffing, and maintenance environment may differ from the development setting. These are not secondary implementation details. They determine whether inputs are available, whether the model remains calibrated, and whether an output can lead to care. Transfer therefore requires local problem definition, external and local validation, workflow testing, and monitoring rather than simple software installation.

What does the digital divide mean for global health AI?

The divide includes more than internet access. It includes reliable electricity, affordable devices and data, compatible information systems, local-language interfaces, technical staff, procurement capacity, and the authority to maintain or replace a system. A cloud-dependent product may exclude facilities with intermittent connectivity, while a smartphone-only workflow may exclude workers or patients who use basic phones. Design should begin with measured constraints at the intended sites.

What is data colonialism in health AI?

Data colonialism describes extractive arrangements in which data and analytic value move away from source communities while decision authority, authorship, infrastructure, and benefits remain elsewhere. A fair partnership defines local governance, permitted uses, access, benefit sharing, authorship, capacity building, return of results, and exit conditions before data collection. National control alone is not sufficient if affected communities and local institutions lack meaningful participation.

How should AI be designed for resource-constrained settings?

Use the simplest technology that can solve the defined problem. Offline-first operation, low-power hardware, graceful failure, local language support, repairability, transparent rules, and compatibility with existing reporting systems may matter more than marginal benchmark gains. The system should preserve a manual fallback and should be maintainable with locally available skills and financing. A pilot should measure service delivery and equity outcomes, not only algorithm accuracy.

How can bias widen global health inequity?

A model can underperform when the training data omit relevant populations, devices, disease patterns, or care pathways. Even equal model performance can produce unequal benefit when confirmatory testing, referral, treatment, or connectivity is unavailable. Equity evaluation should therefore examine access, data quality, error rates, action completion, and outcomes across relevant groups and locations. Mitigation may require redesigning the service pathway rather than only changing the model.

What makes a global health AI example credible?

The evidence should name the exact product or model version, setting, population, comparator, reference standard, workflow, and endpoint. Deployment counts, screening costs, workload effects, and patient outcomes should not be transferred from a vendor announcement or a different study. A technically successful pilot becomes a credible implementation case only when the supporting source and the claimed result match and the limits are explicit.

Discussion Questions

  1. Digital divide: How can we ensure AI does not widen health inequities between and within countries? What role should high-income countries play?

  2. Context appropriateness: Should we prioritize simpler, locally-developed solutions over sophisticated but externally-dependent ones? What trade-offs are acceptable?

  3. Capacity building: What is the best way to build sustainable AI capacity in LMICs? Short-term training? Long-term fellowships? Technology transfer? Building from scratch?

  4. Data governance: How can we ensure LMIC data is not exploited by high-income countries or corporations? What does equitable data partnership look like?

  5. Algorithmic fairness: Whose responsibility is it to ensure AI works for all populations? Developers? Regulators? International organizations?

  6. Innovation models: Should LMICs focus on adapting existing AI tools or developing their own? What are the pros and cons of each approach?

  7. Sustainable funding: How can LMICs sustainably fund AI infrastructure and capacity when facing competing health priorities? Should international donors support AI?

  8. Evaluation standards: Should we apply the same evaluation standards to AI in LMICs as high-income countries? Or should “good enough” be acceptable when the alternative is no access to specialists?

Further Resources

Books and Reports

Global health and equity: - WHO (2021). “Ethics and Governance of AI for Health” - Framework for responsible AI in global health - Abebe et al. (2021). “Roles for Computing in Social Change” - Critical perspectives on technology and development - Madianou (2019). “Technocolonialism: Digital Innovation and Data Practices in the Humanitarian Response to Refugee Crises”

AI in resource-limited settings: - Wahl et al. (2018). “Artificial Intelligence (AI) and Global Health” - Lancet overview - Schwalbe & Wahl (2020). “AI and the Future of Global Health”

Key Papers

Infrastructure and access: - Wahl et al. (2018). “Artificial intelligence (AI) and global health: how can AI contribute to health in resource-poor settings?” BMJ Global Health - Shaw et al. (2023). “Artificial intelligence and health equity in global health.” The Lancet Global Health

Algorithmic fairness: - Obermeyer et al. (2019). “Dissecting racial bias in an algorithm used to manage the health of populations.” Science - Sjoding et al. (2020). “Racial bias in pulse oximetry measurement.” NEJM - Gichoya et al. (2022). “AI recognition of patient race in medical imaging.” Lancet Digital Health - Agarwal et al. (2018). “A reductions approach to fair classification.” ICML - Fairness-aware training

Successful implementations: - Gulshan et al. (2016). “Development and validation of a deep learning algorithm for detection of diabetic retinopathy.” JAMA - Yang et al. (2020). “Deep learning for smartphone-based malaria parasite detection in thick blood smears.” IEEE J Biomed Health Inform - Murphy et al. (2020). “Computer aided detection of tuberculosis on chest radiographs: An evaluation of the CAD4TB v6 system.” Scientific Reports

Data governance: - Hummel et al. (2021). “Data sovereignty in Africa.” Information & Communication Technologies in Africa - Principles for Digital Development: https://digitalprinciples.org

Tools and Frameworks

Fairness tools: - Fairlearn (Microsoft): Fairness assessment and mitigation - https://fairlearn.org - AI Fairness 360 (IBM): Full suite of fairness metrics and algorithms - https://github.com/Trusted-AI/AIF360 - What-If Tool (Google): Visual fairness investigation - https://pair-code.github.io/what-if-tool

Offline-first tools: - TensorFlow Lite: On-device ML - https://www.tensorflow.org/lite - ML Kit: Mobile ML SDK - https://developers.google.com/ml-kit - ONNX Runtime: Cross-platform inference - https://onnxruntime.ai

Population health mapping: - Google Population Dynamics Foundation Model (PDFM): AI model predicting local health and demographic patterns in areas lacking comprehensive data. Combines satellite imagery, search trends, and geographic data through graph neural networks. Freely available on GitHub with a one-minute Colab notebook for resource-limited teams with unreliable connectivity. Applications include measles vulnerability mapping, dengue forecasting, and respiratory disease tracking.

Open health datasets (global): - Global Health Data Exchange: http://ghdx.healthdata.org - WHO Global Health Observatory: https://www.who.int/data/gho - Digital Health Atlas: https://digitalhealthatlas.org

Training Programs

Capacity building: - Data Science Africa: Annual workshop + year-round training - http://www.datascienceafrica.org - AI4D Africa: Fellowships and training - https://ai4d.ai - DeepLearning.AI: Free online courses (available globally) - https://www.deeplearning.ai - Fast.ai: Practical deep learning for coders - https://www.fast.ai

Global health AI: - AI for Global Health (Coursera): Overview course - Digital Health & AI (WHO): Online training modules

Organizations and Networks

Advocacy and research: - AI4D Network: African AI research network - https://ai4d.ai - Global Partnership on AI (GPAI): International AI governance - https://gpai.ai - Research ICT Africa: African ICT policy research - https://researchictafrica.net

Implementation partners: - Digital Square (PATH): Digital health in LMICs - https://digitalsquare.org - Médecins Sans Frontières: AI in humanitarian settings - https://www.msf.org

Key Resources for Practitioners

  • WHO AI Ethics Toolkit: https://www.who.int/publications/i/item/9789240029200
  • Context-Appropriate AI Design Checklist: See exercise above
  • Data Sharing Agreement Template: See example above
  • Fairness Audit Template: See code example above

Looking Ahead

In the next chapter, Policy and Governance, we’ll examine:

  • Policy frameworks for AI in public health
  • Regulatory approaches across countries
  • Governance structures for health AI
  • Accountability mechanisms
  • International cooperation on health AI standards

The equity principles covered in this chapter are essential foundations for the policy discussions ahead.