Future events and outcomes could have a significant impact on our decision-making today. The ability to accurately forecast future outcomes has been shown to add value to many economic agents, such as policymakers, market participants and financial managers. Given the value of accurate forecasts, the development and testing of forecasting tools in the literature are vast. In this light, identifying optimal models to forecast financial and economic variables has been a point of great consideration in literature. Decades ago, a central challenge in forecasting was a lack of data availability. Today, a key challenge is the significant volume of data available and the structuring of models that could accurately work with big data. The Big Data revolution has transformed the modern world and is an important data mining topic that spans across fields. Data mining forms the basis for AI and machine learning (ML) and works together to answer questions, prove hypotheses and give insight into the behaviour of time series. Time series forecasting is part of ML and it has evolved over time from simple linear methods to non-linear methods and complex deep learning (DL) methods, showing a shift from supervised to unsupervised learning. The main aim of this chapter is to establish the evolution and future of time series forecasting through the lens of AI. The findings of this research show that due to the data revolution, challenges and opportunities exist for research, education, current research fields and new research fields. Therefore, looking at the future—the lens we use should focus on a multidisciplinary approach to solving complex problems in time series forecasting.