Dr. Aditi Jhaveri

Embracing Global Englishes in Language Education: Leveraging ChatGPT for Linguistic Diversity 

This presentation showcases the use of ChatGPT as a pedagogical tool to introduce students to a variety of English forms, employing a functional approach to language education grounded in context. The primary objective of the study was to nurture students' appreciation of linguistic diversity by using ChatGPT-generated Singlish, Hinglish, and Chinglish texts in the classroom. By leveraging the model's capabilities, students were able to examine the linguistic features and communicative styles of different varieties of English. Students actively interacted with ChatGPT, posing questions and receiving responses that showed the unique characteristics of each language form in terms of code-mixing, use of loanwords, and pragmatic norms. But students reflected that the training data used to develop ChatGPT might be skewed towards standard English as it may not have been exposed to a wide range of texts on specific English variants. They also felt that it may not accurately capture the unique linguistic features that define a particular English variety. Consequently, the talk ends by suggesting that large language models (LLMs) expand their training data to include a diverse range of texts and language samples from different regions, while also incorporating annotated data that explicitly identifies and highlights the linguistic characteristics of specific English variants. This can enable AI-enhanced language education to move towards the goal of decolonizing English by acknowledging and valuing the richness of non-standard English varieties.