THE EFFECTIVENESS OF NEURAL NETWORKS AND LARGE LANGUAGE MODELS (LLMs) IN DEVELOPING STUDENTS' ACADEMIC WRITING SKILLS

Abstract

This article examines the empirical effectiveness of utilizing Neural Networks and Large Language Models (LLMs) to enhance academic writing skills among students in higher and technical education. The study addresses the pedagogical transition from traditional summative grading to AI-driven formative feedback loops. Through a methodical analysis of scaffolding techniques provided by generative AI tools, the paper highlights how LLMs improve structural coherence, lexical diversity, and syntactic complexity in student essays. Furthermore, it discusses the critical challenges of academic integrity and suggests a process-oriented assessment model to optimize human-AI collaboration in linguodidactics.