Abstract
This study explores the integration of an innovative AI-driven Corpus (AIDC) tool, CorpusChat, in enhancing discipline-specific academic writing skills. CorpusChat combines Corpus Linguistics and Generative AI (GenAI) to simplify corpus searches and provide authentic, data-driven language insights without requiring technical expertise. Using a variety of discipline-specific corpora (e.g., Arts and Humanities, Dentistry, Science, and Medicine) analyzed through AI algorithms, undergraduates at the University of Hong Kong (HKU) have a new learning opportunity to engage with discipline-specific chatbots on CorpusChat to understand and analyze language patterns within discipline-specific corpora. On CorpusChat, specialized chatbots are empowered with the capacity to analyze language patterns within discipline-specific corpora. These chatbots target at the needs of discipline-specific academic writing. The unique capabilities of CorpusChat can address the concerns of GenAI such as accuracy, reliability, and hallucinations.
This team presentation first explains the protocol of and rationale for the development of CorpusChat. The AIDC tool architecture is built on the grounds of merging Corpus Linguistics with Generative AI. We will highlight the benefits and innovations of such architecture and the different features and functions of CorpusChat. In the second part of the presentation, we will demonstrate how CorpusChat is used to support the teaching of academic writing in a first-year Academic Communication in English (ACE) class and an English-in-the-Discipline class for second year students from the Faculty of Arts. The demonstration will take the audience through the journey how GenAI and corpus linguistics can bridge the gap by offering more innovative English language learning opportunities for students in enhancing their academic writing skills. The presentation will end with some sharing of perspectives from a student partner who were actively involved in using CorpusChat in his English class and co-compiling one of the corpora for the CorpusChat.