Beyond Adaptivity: A Modular NLP-Driven Framework for Dynamic Learner Profiling and Generative Educational Content

Generative Artificial Intelligence (GAI) is rapidly transforming the education landscape, enabling both teachers and students to benefit from new forms of personalization. However, systems that aim to be adaptive are often tied to, and therefore limited by, predefined and static parameters, which limit their adaptability to the user. This article presents a modular framework that uses natural language processing (NLP) technologies to extract significant features from student-written contributions to build dynamic profiles that model their abilities and needs. These generated profiles are then used to guide Large Language Models (LLMs) in producing learning content that can be deemed adaptive and personalized. The proposed system integrates and aims to offer a real benefit to both students and teachers. The article will focus on the description of the designed pipeline, its applications in different educational contexts, the evaluation methodology, and, finally, the resulting ethical safeguards.