ChatGPT4 for Classroom Task Adaptation: An Experimental Study of Material Customisation through Student Data Sets
Aleena Khan, The English and Foreign Languages University, Hyderabad (India)
Abstract
For tech integration in language classrooms, ChatGPT4 is a widely used tool in the Indian context. Although prompt engineering as a teacher training element has been explored by previous studies, the current paper explores the operationalization of the LLM to cater to the diversity of the state level schools in India in terms of the cognitive ability and English language competence that stands at varying levels amongst the students of the same classroom. The experimental approach tests if students’ writings can be used as data sets to contextualize prompts asking for fine-tuned adaptations of classroom tasks for a more diverse classroom. In line with the SDG 4 for inclusion in education, the present study explores how ethical and effective use of technology can be ensured through adaptable methods of incorporation instead of context-exclusive tools or software. The study has been conducted on 80 students and 4 teachers for a period of 1 month to understand the feasibility and the potential in this approach to tech integration. This has been done through a qualitative record of teacher insights gathered through group discussions and semi structured interviews, and through student performance in response to their personalized adapted tasks over the period of the intervention.
Keywords |
Tech-integration, ChatGPT4, task adaptation, personalized learning, inclusion |
REFERENCES |
[1] Liu, Pengfei, et al. ‘Pre-Train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing’. ACM Comput. Surv., vol. 55, no. 9, Association for Computing Machinery, Jan. 2023, https://doi.org10.1145/3560815
[2] Ali, Zuraina. ‘Artificial Intelligence (AI): A Review of Its Uses in Language Teaching and Learning’. IOP Conference Series: Materials Science and Engineering, vol. 769, no. 1, IOP Publishing, Feb. 2020, p. 012043, https://doi.org10.1088/1757-899X/769/1/012043.
[3] Radford, A., et al. (2019). Language Models are Unsupervised Multitask Learners. OpenAI. |