Scholarlink

<p>Teaching German as a second language (GSL) in heterogeneous classrooms requires access to authentic and level-appropriate reading materials, yet educators often struggle to find resources that are both pedagogically suitable and legally usable. This paper presents the LATILL platform, an AI-driven tool designed to support teachers in identifying, adapting, and organizing German texts according to the Common European Framework of Reference for Languages (CEFR). The platform integrates a curated corpus of over 20,000 authentic texts, automated CEFR-level classification, and topic tagging, enhanced with generative AI functionalities for text simplification, translation, and visualization. Through iterative, user-centered development and evaluation with more than 40 educators across Germany, Spain, and Ukraine, LATILL demonstrated strong potential for improving lesson planning, differentiation, and classroom adaptability. Empirical testing showed acceptable accuracy in CEFR classification (around 72–73%) and highlighted the value of AI features, while also identifying challenges in text simplification, translation fidelity, and image generation. Successive refinements, informed by teacher feedback and interaction analytics, improved usability, content quality, and pedagogical alignment, leading to increased engagement and sustained adoption. The study concludes that LATILL not only addresses the persistent challenge of sourcing and tailoring texts for diverse learner groups but also establishes a sustainable, teacher-centered infrastructure for integrating AI into language education.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

From CEFR classification to generative AI materials: designing and validating the LATILL platform

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月15日

查看全文
  • Andrea Vázquez-Ingelmo,
  • Denise Löfflad,
  • Alicia García-Holgado,
  • Roberto Therón-Sánchez,
  • Detmar Meurers

摘要

Teaching German as a second language (GSL) in heterogeneous classrooms requires access to authentic and level-appropriate reading materials, yet educators often struggle to find resources that are both pedagogically suitable and legally usable. This paper presents the LATILL platform, an AI-driven tool designed to support teachers in identifying, adapting, and organizing German texts according to the Common European Framework of Reference for Languages (CEFR). The platform integrates a curated corpus of over 20,000 authentic texts, automated CEFR-level classification, and topic tagging, enhanced with generative AI functionalities for text simplification, translation, and visualization. Through iterative, user-centered development and evaluation with more than 40 educators across Germany, Spain, and Ukraine, LATILL demonstrated strong potential for improving lesson planning, differentiation, and classroom adaptability. Empirical testing showed acceptable accuracy in CEFR classification (around 72–73%) and highlighted the value of AI features, while also identifying challenges in text simplification, translation fidelity, and image generation. Successive refinements, informed by teacher feedback and interaction analytics, improved usability, content quality, and pedagogical alignment, leading to increased engagement and sustained adoption. The study concludes that LATILL not only addresses the persistent challenge of sourcing and tailoring texts for diverse learner groups but also establishes a sustainable, teacher-centered infrastructure for integrating AI into language education.