Year: 2026
Type: article
Abstract: Generative AI has fundamentally reshaped many research practices within academic institutions. One area experiencing substantial change is qualitative discourse analysis within the sociolinguistic paradigm. Adopting a critical realist-sociotechnical perspective, this study positions large language models (LLMs) not as replacements for human analysts but as components of emergent analytical systems where human-AI collaboration yields novel capabilities. The study examines two discourse analysis excerpts from the author’s pre-GenAI research, subjecting them to zero-shot prompting using GPT-4o and Gemini 2.5 Flash. By comparing the models’ outputs with the original human analyses, the research identifies opportunities for genuine synergy between analysts and AI. The findings show that LLMs, particularly ChatGPT, produce sophisticated interactional sociolinguistic analyses and excel at maintaining theoretical consistency, though tend to lack the contextual depth and cultural specificity that human analysts can provide. The paper advocates for a synergistic approach where LLMs augment rather than replace discourse analytic practice, proposing task-appropriate divisions of labour for researchers who want to integrate GenAI into their workflows. As qualitative research grapples with AI integration, this study suggests that the human-AI system represents a site of genuine analytical emergence which can transcend the capabilities of either component alone.
Author: Jay M. Woodhams
Language: English
Subfield: General Social Sciences
Field: Social Sciences
Domain: Social Sciences
Open Access status: hybrid
Estimated APC paid: $2,000
OpenAlex ID: W7128613272