Exploring User Creativity and Ethical Decision Making in Generative AI: A Transformer-Based Semantic and Network Analysis of Reddit Discussions on ChatGPT

Satam, Gaurav Sadanand (2025) Exploring User Creativity and Ethical Decision Making in Generative AI: A Transformer-Based Semantic and Network Analysis of Reddit Discussions on ChatGPT. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (965kB) | Preview
Preview	PDF (Configuration Manual) Download (1MB) | Preview

Abstract

The creation of generative AI applications, such as ChatGPT, at a very fast pace has changed the balance of interaction between individuals and has already begun to redefine creativity and ethical choice in daily life. The research questions the use of ChatGPT in teaching Reddit communities by students and the inventiveness and ethical behaviour dynamic. The study measures the emotional, ethical, and creative profile of 177,485 posts on Reddit with the help of the How Students Use ChatGPT Filtered Reddit Dataset and an advanced transformer-based pipeline, which combines RoBERTa sentiment analysis, BART-based zero-shot ethical stance classification, and Sentence-BERT semantic embeddings. It is revealed that the student discourse is mostly neutral to positive (neutral 50.5%, positive 31.7%, negative 17.8%), and the majority of the conversations involve the description of pragmatic and informational ChatGPT uses, instead of polarised responses. The prosocial posts (53.3%), and the neutral ones (23.4%), and the posts showing potentially harmful or risky usage (12.8%), are identified as the ethical stance modelling demonstrates that more than half of the posts are prosocial (53.3%), and another significant proportion (23.4 percent) is the neutral ones. Creativity scores show that discussions are more informative than radically innovative with zero-shot labels being dominated by the informative posts (78.1%) and very few posts being highly creative (1.6%), which may suggest that students use ChatGPT more as an intellectual assistant than as the source of original invention. Combining these composite scores with a lightweight co-occurrence network, which identifies central tags like Conversation and College as central discourse nodes, the study has provided a multidimensional analytical paradigm that will contribute to the existing knowledge regarding the impact that generative AI is having on student creativity, ethics, and community norms in the real-world online context.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Muntean, Cristina Hava UNSPECIFIED
Subjects:	B Philosophy. Psychology. Religion > BJ Ethics P Language and Literature > P Philology. Linguistics > Computational linguistics. Natural language processing Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence > Generative artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence > Generative artificial intelligence Z Bibliography. Library Science. Information Resources > ZA Information resources > ZA4150 Computer Network Resources > The Internet > World Wide Web > Websites > Online social networks T Technology > TK Electrical engineering. Electronics. Nuclear engineering > Telecommunications > The Internet > World Wide Web > Websites > Online social networks
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	09 Sep 2026 08:34
Last Modified:	09 Sep 2026 08:40
URI:	https://norma.ncirl.ie/id/eprint/9905

Actions (login required)

View Item