AI in Healthcare — Africa Continental Dashboard
Contextual and seven-dimension risk analysis of clinical-AI adoption across 53 African countries, spanning Advanced, Limited, Planned and None tiers. Regulatory approval, data sensitivity, human oversight, validation rigor and data sovereignty are each measured and aggregated into an overall risk score.
1 · Executive summary
- Only 4 of 53 African countries (Egypt, Nigeria, South Africa, Kenya) have an advanced clinical-AI deployment footprint.
- 7 countries (Morocco, Rwanda, Ethiopia, Mauritius, Ghana, Uganda, Guinea) are at the 'Limited' tier, running one to two documented AI models.
- 4 countries are actively planning AI adoption; 38 have no documented deployments.
- A total of 29 AI models are catalogued across the continent.
- The composite-risk model flags 42 countries at Critical risk — primarily because no deployed AI is documented. Morocco has the lowest composite risk (25/100).
- Regulatory approval is largely 'UNKNOWN' except for a handful of models (Lunit INSIGHT, Qure.ai, CAD4TB) that carry FDA or CE mark; validation rigor is dominated by retrospective / pilot studies rather than prospective trials.
2 · Adoption landscape
How many African countries have reached each adoption level, and how is that distributed across regions?
3 · Risk measurement
Seven risk dimensions are computed per country; each is normalised to 0–100 (higher = riskier).
Risk-dimension glossary
- Adoption Gap — Distance from the 'Advanced' adoption tier. Countries with no documented AI deployment (None/Planned) score highest.
- Regulatory — Share of deployed models that lack a recognised regulatory approval (FDA, CE mark, ministry-of-health rollout, etc.).
- Data Sensitivity — Average sensitivity level of the patient data processed by the deployed models. High-sensitivity data (imaging, EHRs) raises privacy & consent risk.
- Human Oversight — Inverse of the measured human oversight level. Lower physician review increases the risk of AI-driven clinical errors.
- Validation Rigor — Inverse of the validation method's rigor — lack of prospective / multi-site validation implies weak clinical evidence.
- Data Sovereignty — Share of deployed models storing data abroad or in foreign cloud infrastructure — a privacy and jurisdictional risk.
- Transparency — Share of contextual fields marked 'unclear' / 'undocumented'. Opaque reporting blocks audit and post-market surveillance.
4 · Country summary table
Sorted by overall risk.
| Country | Region | Adoption | Models | Reg. risk | Sens. risk | Trans. risk | Overall | Tier |
|---|---|---|---|---|---|---|---|---|
| Mauritania | West Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Niger | West Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Togo | West Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Angola | Central Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Central African Republic | Central Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Chad | Central Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Republic of the Congo | Central Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Democratic Republic of the Congo | Central Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Equatorial Guinea | Central Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Gabon | Central Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| São Tomé and Príncipe | Central Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Burundi | East Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Comoros | East Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Djibouti | East Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Madagascar | East Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Malawi | East Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Mozambique | East Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Seychelles | East Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| South Sudan | East Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Tanzania | East Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Zambia | East Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Zimbabwe | East Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Botswana | Southern Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Lesotho | Southern Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Namibia | Southern Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Sierra Leone | West Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Eswatini | Southern Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Libya | North Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Liberia | West Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Guinea-Bissau | West Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Gambia | West Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Côte d’Ivoire | West Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Cabo Verde | West Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Burkina Faso | West Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Benin | West Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Sudan | North Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Mali | West Africa | None | 0 | 100 | 100 | 100 | 100 | Critical |
| Cameroon | Central Africa | Planned | 1 | 100 | 100 | 100 | 97 | Critical |
| Algeria | North Africa | Planned | 1 | 100 | 100 | 100 | 97 | Critical |
| Tunisia | North Africa | None | 0 | 100 | 100 | 86 | 94 | Critical |
| Senegal | West Africa | Planned | 1 | 100 | 100 | 86 | 88 | Critical |
| Somalia | East Africa | Planned | 1 | 100 | 100 | 86 | 88 | Critical |
| Mauritius | East Africa | Limited | 2 | 100 | 50 | 57 | 77 | High |
| Guinea | West Africa | Limited | 1 | 100 | 100 | 43 | 68 | High |
| Ghana | West Africa | Limited | 1 | 100 | 100 | 29 | 67 | High |
| Ethiopia | East Africa | Limited | 2 | 100 | 100 | 29 | 63 | High |
| Uganda | East Africa | Limited | 1 | 100 | 100 | 43 | 62 | High |
| Nigeria | West Africa | Advanced | 5 | 100 | 62 | 57 | 51 | Moderate |
| Egypt | North Africa | Advanced | 1 | 100 | 100 | 43 | 48 | Moderate |
| Kenya | East Africa | Advanced | 4 | 100 | 50 | 57 | 46 | Moderate |
| South Africa | Southern Africa | Advanced | 4 | 40 | 100 | 43 | 44 | Moderate |
| Rwanda | East Africa | Limited | 2 | 50 | 100 | 29 | 33 | Low |
| Morocco | North Africa | Limited | 2 | 50 | 50 | 14 | 25 | Low |
5 · Country-by-country deep dive
Filter by tier / adoption / region, or search by name.
Egypt Advanced ModerateRisk 48
Models: Vezeeta, Manentia, AI breat cancer detection platform, National AI Remote Diagnostics Platform, Vision Transformer (ViT), and Llama3-OpenBioLLM-70B
Clinical focus: - Vezeeta : patient assitance, telehealth - Manentia AI : General clinical care, telehealth - AI breast cancer detection platform : oncology/ radiology - National AI Remote Diagnostics Platform : General medicine / public health / emerge…
Regulatory: - Vezeeta : UNKNOWN - Manentia AI : UNKNOWN - AI breast cancer detection platform : UNKNOWN - National AI Remote Diagnostics Platform : UNKNOWN - Vision Transformer (ViT) : UNKNOWN - Llama3-…
Oversight: - Vezeeta : high (Human decision is primary) - Manentia AI : high (physicians have to review all ouputs) - AI breast cancer detection platform : high (physicians have to review diagnosis before informaing patients) - N…
Real-world use: - Vezeeta : fully functional service with millions of bookings and active users - Manentia AI : unclear - AI breast cancer detection platform : launched and integrated into Baheya Hospital workflows (Haram and Sheikh Z…
AI policy: national AI strategy since 2021, but still developping policies
Nigeria Advanced ModerateRisk 51
Models: - AwaDoc : Symptom analysis + triage chatbot via WhatsApp - Clafiya : Predictive analytics for disease surveillance - SmartMRS : AI‑assisted EHR/diagnosis + imaging analysis - Drug Insights EMDEX AI : web app for healthcare professionals nationally to support…
Clinical focus: - AwaDoc : Primary care support & health information / triage - Clafiya : Public health & preventive care support - SmartMRS : Clinical decision support & patient data management - Drug Insights : Clinical decision support / health informa…
Regulatory: - AwaDoc : UNKNOWN - Clafiya : UNKNOWN - SmartMRS : UNKNOWN - Drug Insights : UNKNOWN - Intron Health’s Speech‑to‑Text AI : UNKNOWN
Oversight: - AwaDoc : N/A - Clafiya : High - Content and guidance are curated and supervised by health professionals or NGOs before dissemination. - SmartMRS : High- All AI outputs are reviewed by clinicians before final decisions…
Real-world use: - AwaDoc : used by nearly 30 000 people - Clafiya : also used - SmartMRS : Unclear - Drug Insights : Unclear - Intron Health’s Speech‑to‑Text AI : actively deployed in numerous Nigerian hospitals including University Co…
AI policy: national AI strategy --> present but still being developped
South Africa Advanced ModerateRisk 44
Models: Envisionit Deep AI (AI radiology platform for chest X-rays) Lunit INSIGHT (Deep learning models for: Chest X-ray analysis, Breast cancer detection) Qure.ai (AI imaging tools for: TB detection (qXR) and emergency findings (qER)) PathAI (AI models analyzing bio…
Clinical focus: Envisionit Deep AI : Medical imaging (radiology) Lunit INSIGHT : Oncology + radiology Qure.ai : Radiology / emergency diagnostics, PathAI : Cancer diagnosis (histopathology) CAD4TB : Infectious disease (TB screening)
Regulatory: Envisionit Deep AI : UNKNOWN Lunit INSIGHT : FDA + CE mark Qure.ai : FDA + CE mark PathAI : UNKNOWN CAD4TB : CE mark
Oversight: Envisionit Deep AI : High – radiologists/hospital staff always review AI outputs before decisions Lunit INSIGHT :High – radiologists/hospital staff always review AI outputs before decisions Qure ai : High – radiologist …
Real-world use: Envisionit Deep AI YES - Deployed in South African hospitals and ICUs — used in a 700‑bed hospital in the Northern Cape province with limited radiologists to help identify pathologies like TB, COVID‑19 pneumonia, etc. d…
AI policy: AI digital health policies
Kenya Advanced ModerateRisk 46
Models: AntiMicro.ai : tool that predicts the likelihood that a bacterial organism will be resistant to specific antibiotics AI Consult (GPT‑4o LLM) : large language model‑powered clinical decision support tool TRIM AI : Triage SMS‑based maternal health messages to d…
Clinical focus: AntiMicro.ai : Infectious Diseases / Antimicrobial Stewardship AI Consult (GPT‑4o LLM) : Primary Care / General Clinical Decision Support TRIM AI : Maternal Health / Obstetrics Tabibu health : General health information
Regulatory: AntiMicro.ai : UNCLEAR AI Consult (GPT‑4o LLM) :UNCLEAR TRIM AI : UNCLEAR Tabibu health :UNCLEAR
Oversight: AntiMicro.ai : High - clinicians interpret results before prescribing AI Consult (GPT‑4o LLM) : High - doctors remain decision-makers TRIM AI : High - under strict oversight policies Tabibu health : High- clinicians val…
Real-world use: AntiMicro.ai : Actually deployed in approximately 15 clinics and Tested on approximately 40,000 real patient visit AI Consult (GPT‑4o LLM) : no public documentation TRIM AI : no public documentation Tabibu health : no …
AI policy: national digital health strategy but limited AI regulation
Morocco Limited LowRisk 25
Models: Diabetes Glycemic Control Risk Model : machine learning model that predicts which diabetic patients are at risk of poor blood sugar control, Da Vinci surgical system : robotic-assisted surgery platofrom, backed by AI to enhance precision, safety, and ergonomi…
Clinical focus: Diabetes Glycemic Control Risk Model : Endocrinology / Diabetes management Da Vinci surgical system :Surgery / Minimally invasive procedures
Regulatory: Diabetes Glycemic Control Risk Model : unknown Da Vinci surgical system : CE mark
Oversight: Diabetes Glycemic Control Risk Model : High Da Vinci surgical system :High
Real-world use: Diabetes Glycemic Control Risk Model : Hospital/research use in Morocco Da Vinci surgical system : Actual use in Moroccan hospitals (Casablanca, Rabat)
Rwanda Limited LowRisk 33
Models: Babylon AI Triage Tool : telehealth, patient triage tool HealthPulse AI : interprets medical results, improves diagnosis accuracy
Clinical focus: Babylon AI Triage Tool : Telemedicine / triage / primary care decision support HealthPulse AI : Primary care / diagnostics (malaria, rapid tests)
Regulatory: Babylon AI Triage Tool : unclear HealthPulse AI : Ministry of health rollout
Oversight: Babylon AI Triage Tool : High HealthPulse AI : High
Real-world use: Babylon AI Triage Tool : Used daily for remote consultations and triage across Rwanda HealthPulse AI : unclear
Ethiopia Limited HighRisk 63
Models: Bego healthcare : AI‑powered electronic medical record and management platform that includes AI assistance for clinical decision support RadiSen’s AXIR‑CX AI model : A deep learning‑based AI model for detecting TB, pneumonia, and other lung conditions from d…
Clinical focus: Bego healthcare : digital health/ decision support RadiSen’s AXIR‑CX AI model : radiology/oncology
Regulatory: Bego healthcare : unknown RadiSen’s AXIR‑CX AI model : unknown
Oversight: Bego healthcare : High- physicians review all AI output and take decisions RadiSen’s AXIR‑CX AI model : High-physicians review all AI output and take decisions
Real-world use: Bego healthcare : used in multiple clinics RadiSen’s AXIR‑CX AI model : used at Yekatit 12 Hospital (public health)
Mauritius Limited HighRisk 77
Models: DRRIYA : telehealth platform (health assistant tool) --> powered by R.I.Y.A‑AIx‑Lite‑2.0.4‑V7 (LLM tool) RIYA‑AIX‑V2 (clinical tool to handle symptom interpretation and health‑related questions)
Clinical focus: DRRIYA : general health support --> " provides general information only and is not a substitute for professional medical advice, diagnosis, or treatment"
Regulatory: DRRIYA : UNKNOWN
Oversight: DRRIYA : The wording on DRRIYA’s site emphasizes that the AI provides guidance and information, not clinical decision‑making power : "Medical Disclaimer This AI medical assistant provides general information only and is…
Real-world use: DRRIYA : available for the whole country, but no evidence of use
Ghana Limited HighRisk 67
Models: minoHealth.ai radiology AI systems: Deep learning for diagnosing pleural effusion and cardiomegaly from chest X‑rays
Clinical focus: Radiology / Diagnostic imaging
Regulatory: no public documentation
Oversight: High
Real-world use: Assisting chest radiograph interpretation for conditions like cardiomegaly and effusion
Uganda Limited HighRisk 62
Models: Mak ocular : analyzes microscope images to detect diseases faster and more accurately
Clinical focus: Mak ocular : Diagnostics (malaria, TB, cervical cancer)
Regulatory: Mak ocular : unclear
Oversight: Mak ocular : High
Real-world use: Mak ocular : unclear-still pilot
Guinea Limited HighRisk 68
Models: KÔLÈP ÒNO‑AI : AI-assisted rapid detection of malaria and tuberculosis (≈2 minutes)
Clinical focus: Diagnostics (malaria and tuberculosis detection)
Regulatory: unknown
Oversight: high
Real-world use: not guaranteed
Somalia Planned CriticalRisk 88
Models: Deep Learning Pneumonia Detection Model (not deployed yet
Senegal Planned CriticalRisk 88
Models: Kera Health : COMING SOON !!!
Algeria Planned CriticalRisk 97
Models: APC-GNN++ (still in research stage)
Cameroon Planned CriticalRisk 97
Models: DigiCare, SophiA