This research aims to investigate the skills and scope of the AI text generators, ChatGPTand GPT-3.5 in generating academic writing assignments in a social science course to evaluate how much they can meet the assessment criteria of the relevant assignment. Student participants will be asked to use the either tool to complete two different written assignments and they will be interviewed to examine how they utilize the tool as reading and writing aid. And asking questions informed by the following overarching research questions:
1. To what extent is a ChatGPT or GPT-3.5 generated text successful based on the criteria of the relevant writing assignment?
2. Based on their experience in writing with ChatGPT, to what degree do students feel involved in completing the assignment?
3. How should the assessment and pedagogy in writing be revised to minimize the negative implications of tools like ChatGPT and GPT-3.5 in the learning process of academic writing?
Based on the data collected, the research aims to create a set of instruments and suggestions about the types of alternative assessments that also integrate themselves to this new reality. Colloboration of students from computer science and social sciences departments will be the best part of this research to be in good understanding of the technical abilities of the AI tools and discovering the useful skills and stages of academic writing.
About Project Supervisors
Ilkem Kayican Dipcin, Sabanci University- School of Languages, ilkem.kayican@sabanciuniv.edu