top of page
Corbin Donnelly
Keynote Speakers
Prof. Serge Bibauw
Assistant Professor at University of Louvain
Serge Bibauw is an Assistant Professor in French language teaching at the University of Louvain (UCLouvain, Belgium). He holds a Master's degree in Romance Languages from UCLouvain and a PhD in Linguistics from KU Leuven (Belgium). He has a long experience as a teacher, teacher trainer and evaluator in French and English in Ecuador. His research focuses on the evaluation of the effectiveness of conversational AI and dialogue-based CALL for the development of L2 proficiency, particularly in terms of vocabulary and fluency development.
Looking back, looking forward. From dialogue-based CALL to LLM-based chatbots for language learning
While research on applications of dialogue systems and chatbots to language education has a long history, the large language models (LLM) revolution has suddenly brought a new wave of possibilities for language learning. This keynote will look back at 40 years of research on dialogue-based computer-assisted language learning (CALL) to identify what still holds up. We will revisit typologies of systems, tasks and effects (Bibauw, François & Desmet, 2019) in the light of generative AI chatbots. While taxonomies have to evolve, we will show that task design and learner guidance are now more decisive than ever.
Looking forward toward LLM-based chatbots, we will summarize the findings from very recent studies on this new era of conversational agents, identifying trends in learners’ and teachers’ perceptions or usage as well as the most promising implementations in terms of learning effects (e.g., Dizon, 2024; Lai & Lee, 2024; Hınız, 2024; Ji, Han & Park, 2024; Shin & Lee, 2024; Zhang, Meng & Ma, 2024), and propose some new avenues for research.
Abstract
Vanessa Carson
Prof. Frazer Heritage
Lecturer at Manchester Metropolitan University
Frazer Heritage is a corpus linguist with a particular interest in the representation of gender and sexuality across different forms of digital media. His primary research interest examines how corpus linguistics can be used to study videogames, newspapers, and social media. Frazer has published several articles, many of which use corpus linguistic methods to explore videogames from different perspectives - these articles can be found in journals such as Game Studies, Discourse Context & Media, and Corpora. Furthermore, their published work includes two monographs – one in 2021 exploring corpus approaches to gender in videogames and a second at the end of 2023 which examines the language used by the incel community. Frazer is also the lead editor of the 2024 textbook "Analysing Representation", which focuses on the use of corpus linguistic methodologies in tandem with discursive analytical frameworks.
A useful synergy or a black box of snake oil?: Applications of AI to the study of ludolinguistics
This keynote addresses the use of generative AI in the study of ludolinguistics, with a particular focus on potential synergies between AI and corpus linguistics. Three possible synergies are conceptualised and then tested using data generated from ChatGPT4. These potential synergies are: i) the use of AI to identify academic literature on ludolinguistics, ii) consulting AI to show trends in the language used within videogames and iii) the use of AI to analyse data from videogames. The findings demonstrate that ChatGPT4 often identifies scholars working in either ludography or linguistics but struggles to identify academics working on ludolinguistics. I also demonstrate that the AI models draw on outdated and unevidenced data to reach conclusions about the language within games. I take statements generated by ChatGPT4 about the language used by men and women in games and compare these to authentic corpus data. Such analysis reveals significant differences in authentic data compared to statements made by ChatGPT4. Finally, I demonstrate a low confidence in ChatGP4’s ability to categorise collocates. Ultimately, I argue that while ChatGPT4 offers some interesting avenues for future research – and the development of this tool presents an exciting prospect for academics, the use of such tools should be approached with extreme caution.
Abstract
.png)
bottom of page