Generative AI and its dilemmas: exploring AI from a translanguaging perspective

Abstract Technological advances in generative artificial intelligence (GenAI) have produced multimodal and multilingual models that have the potential to achieve a long-sought educational goal: increasing equity and access for linguistically diverse learners. However, despite their capability to receive adaptive inputs combined with their proficiency in multiple languages and modalities, GenAI systems are designed to operate in the realm of statistical probability, standardizing what is already prevalent in society and under-representing minority voices. In this regard, the possibilities and the limitations of GenAI collide, resulting in instructional dilemmas regarding whether to invest effort into opening space for greater social equality or conveniently use the technology to maintain the social inequality of the status quo. Drawing on a translanguaging perspective, we introduce and illustrate the three dilemmas of creativity-standardization, inclusion-fixity, and meaning-form, which educators will inevitably encounter when integrating AI into their classrooms. We further discuss how this new and increasingly powerful technology can be understood and addressed by highlighting translanguaging as an analytical perspective to understanding and resolving the three dilemmas its use entails.