This chapter examines the operational dimension of AI Literacy, focusing on how modern AI systems learn from data and are applied in various contexts. It begins by contextualising the historical shift from the symbolic, instruction-based paradigm (GOFAI) to the now-dominant subsymbolic, data-driven approach of machine learning (ML). The core mechanisms of ML are explained, detailing its main types: supervised, unsupervised, and reinforcement learning. This chapter then provides an in-depth analysis of the current state-of-the-art, explaining the Transformer architecture that powers Large Language Models (LLMs) and Generative AI. Following this technical exposition, a typology of AI applications is presented, categorised by user profiles (programmers, professionals, citizens) and with a specific focus on their use in education. To translate these complex concepts into practice, this chapter concludes by proposing two pedagogical methods: using the trial-and-error technique to teach the principles of classification and employing gamification to help learners understand the processes of estimation and prediction in AI.