Improving Gen-AI’s language interpretation through cognitive linguistics

Researchers are interested in how AI solutions can be critically and effectively integrated into educational settings. This research explores the application of two cognitive linguistic theories to enrich Gen AI’s understanding of PSTs’ input language. It delves into the complexities and nuances of context-dependent teaching connotations, bridging the gap between human language and AI language processing capabilities. It will focus on how Conceptual Metaphor Theory and Mental Spaces Theory can be programmed into AI models to improve their accuracy and contextual understanding. This study employs a qualitative research design, utilising focus group interviews to explore the research questions. Focus groups conducted were group one, 3 participants (1 at the beginning of the program and 2 in the middle of the program) and group two, 4 participants (all at the beginning of the program). This approach was selected to provide deeper insights and uncover nuances that might not be revealed through other data collection methods. The focus group students believe AI-generated literacy tasks are generally accepted as meaningful but only as general ideas and frameworks, not as fully developed tasks, and English teachers were able to more precisely notice shortcomings of the suggested Literacy tasks as the tasks are KLA-specific. The focus group students all found that numeracy was more challenging than Literacy and wanted to use AI to assist in providing information that could be used to develop engaging numeracy activities.