The COVID-19 pandemic has highlighted the vulnerabilities of healthcare systems worldwide. Artificial Intelligence (AI) has emerged as a critical tool in the global response to the pandemic, enabling infection detection, forecasting, response planning, risk assessment, and patient prioritization, among other tasks. This paper presents an in-depth analysis of AI-based systems for COVID-19, focusing on two primary tasks: spatio-temporal modeling and forecasting, and modeling the impact on diverse populations. To conduct this analysis, a comprehensive literature survey was performed using IEEE Xplore, PubMed, and Google Scholar, covering articles published between 2019 and September 2021. A total of 68 research papers were shortlisted from the initial pool of 156, based on their relevance to the scope of this review. In the domain of spatio-temporal modeling and forecasting, various AI models have been employed to predict the trajectory of the pandemic, aiding in resource allocation, policy-making, and containment strategies. Deep learning models such as Long Short Term Memory (LSTM), Convolutional LSTM (ConvLSTM), and Bidirectional LSTM (Bi-LSTM) have shown promising results in forecasting infection cases. Additionally, statistical approaches, time series models, and dynamic Susceptible Exposed Infectious Recovered (SEIR) models have been used to analyze the spread of the virus. In the context of diverse populations, AI models have been utilized to understand the impact of intrinsic and extrinsic factors on the spread and severity of COVID-19. These factors include medical conditions, age, environmental differences, and social, economic, and political factors. By considering population diversity, AI-based systems can provide insights into differential impacts and inform targeted mitigation strategies. This review highlights the significance of AI systems in combating the COVID-19 pandemic and emphasizes the need for more rigorous categorization of research based on specific AI tasks. It also discusses the potential of AI systems in the near term and emphasizes the importance of learning from our experiences to better prepare for future pandemics. The findings presented in this paper contribute to the growing body of knowledge on AI's role in healthcare and provide valuable insights for researchers, policymakers, and healthcare practitioners.