Abstract
<jats:p>This article examines the convergence of Generative Artificial Intelligence and Learning Analytics in education through a structured literature review. It highlights how large language models expand analytical capabilities by enabling the interpretation of unstructured learner data and supporting automated feedback, personalized learning, and multimodal analysis. At the same time, the study identifies critical risks, including hallucinations, bias, opacity, over-reliance, and threats to academic integrity and privacy. Particular emphasis is placed on the EU AI Act, which classifies many educational AI applications as high-risk and introduces strict requirements for transparency, documentation, and human oversight. The article argues for responsible-by-design adoption, integrating governance frameworks, explainability mechanisms, and AI literacy for educators. It concludes by outlining future research needs related to long-term learning impact, fairness, and institutional AI governance.</jats:p>