Defining artificial intelligence (AI) remains a complex challenge, given its rapidly evolving nature. Nevertheless, one broadly accepted definition—endorsed by the European Union—describes AI as a suite of algorithms that learn from data to make predictions. Since the early 2010s, AI has emerged as a highly effective approach for analyzing, processing, and even generating medical image data. In this chapter, we provide an overview of key AI paradigms—radiomics, deep learning (DL), and foundation models—and examine their current and potential applications in medical image analysis within radiation oncology. We focus on topics, such as image classification, treatment planning, and response assessment, as well as novel strategies involving vision language models (VLMs). Through this exploration, we offer insights into how AI is transforming clinical workflows and shaping future directions in radiation oncology.