Large Language Model-Based Artificial Intelligence in the Language Classroom: Practical Ideas for Teaching

Large Language Models (LLMs) are a powerful type of Artificial Intelligence (AI) that simulates how humans organize language and are able to interpret, predict, and generate text. This allows for contextual understanding of natural human language which enables the LLM to understand conversational human input and respond in a natural manner. Recent examples of this, such as the Generative Pre-Trained Transformer (GPT) model, popularized by OpenAI's web application, ChatGPT, are able to complete an astounding variety of tasks when provided with simple language input. For education, LLMs can alleviate teacher curriculum and grading workloads and even perform specific tasks such as generating creative ideas for activities. Specifically for language learning, LLMs can draw on their immense corpus of language content to generate learner-centric materials to aid teachers in delivering targeted, personalized language instruction. The aim of this paper is to provide the reader with examples of how LLMs can be utilized for materials development, classroom activities, and providing feedback. After giving specific examples and explanations, the paper will conclude with a discussion of how this technology can provide teachers with new innovative ways to streamline the teaching process to focus on learner needs.