A systematic literature review of artificial intelligence methods applied to the Human Epidemic (Covid-19)

The COVID-19 pandemic placed unprecedented pressure on healthcare systems worldwide and accelerated the use of artificial intelligence (AI) in areas such as diagnostics, forecasting, treatment, and disease monitoring. To better understand this trend, we carried out a systematic literature review following PRISMA guidelines. Our review covered studies released between December 2019 and January 2024, drawing from major bibliographic databases and preprint servers. For the search, we applied the keywords (“Artificial Intelligence” OR “Machine Learning” OR “Deep Learning”) AND (“COVID-19” OR “Coronavirus” OR “Pandemic”). Out of approximately 110 retrieved records, 57 studies satisfied the predefined inclusion criteria and were evaluated using the CASP framework. The final body of literature was distributed as follows: disease detection and prediction accounted for about 44% (23 studies), drug and vaccine discovery for 25%, remote healthcare and IoT applications for 19% (10 studies), online social network (OSN) analytics for 4%, and general or multi-modal AI frameworks for 8%. Deep learning models applied to chest CT scans, X-rays, and RT-PCR enhancement frequently reported internal diagnostic accuracies above 90%. Meanwhile, natural language processing and embedding-based OSN methods occasionally identified symptomatic trends several days ahead of official case reports. Despite these advances, common challenges persisted, including heterogeneous datasets, insufficient external validation, and ongoing privacy and ethical concerns. Looking forward, we recommend that future research emphasize multimodal integration of clinical and social media data, establish standardized external benchmarks, adopt explainable AI (XAI) methods, and explore privacy-preserving strategies such as federated learning to strengthen generalizability and promote equitable deployment.