Utilization of artificial intelligence for infectious disease in Africa: a systematic review

Introduction Artificial intelligence is becoming increasingly useful in the management of various diseases across Africa. It is a transformative tool that its use in Africa has been limited due to various factors, including limited digital infrastructure, funding, poor-quality health data, and policy frameworks. This paper aimed to systematically review how AI has been applied across infectious diseases in Africa and explore the challenges and opportunities that exist. Methods This systematic review was conducted using the PRISMA 2020 guidelines across PubMed, AJOL, Scopus, Google Scholar, and ScienceDirect databases. It was registered on PROSPERO with the identification number CRD420251243329. Original peer-reviewed articles in which artificial intelligence such as machine learning or deep learning was used for infectious diseases management, prediction, surveillance, or outbreak prediction conducted in Africa from 2015 to 2025 were included using predefined inclusion criteria. A total of 5050 records were obtained across the databases, with PubMed 2551, ScienceDirect 2000, AJOL 102, Scopus 233, and Google Scholar 164. Data on study characteristics, key findings, and performance metrics of the used AI models were systematically extracted and presented. Results 5006 records were screened, of which 43 eligible studies were included. The most common countries where AI was applied to infectious disease prediction, management, or surveillance were South Africa (n=10), Nigeria (n=8), Uganda (n=6), Ethiopia (n=5), and Tanzania (n=4). They were used for common diseases such as HIV, malaria, COVID-19, and tuberculosis. The most common AI models used in the studies were random forest, support vector machine, artificial neural network, and logistic regression for machine learning models and long short-term memory for deep learning, with reported performance metrics. Conclusion There is increasing use of AI for infectious diseases in Africa. Our study showed common AI types such as random forest, support vector machine, artificial neural network, and logistic regression being used for the prediction, management, and surveillance of common diseases such as HIV, malaria, COVID-19, and tuberculosis. However, it is still underutilized due to factors surrounding policy frameworks, data quality, infrastructure, funding of healthcare systems and research, and governance. There is a need for investments and commitments to address these challenges. PROSPERO ID: CRD420251243329