Co-Creating Meaning: AI-Assisted Coding in Critical Discourse Analysis

Introduction

The integration of generative artificial intelligence (GenAI) into qualitative research has opened new opportunities for enhancing methodological approaches. There is an increasing interest in incorporating GenAI tools for discourse analytical studies as well (DeJeu, 2024). In the context of Critical Discourse Analysis, which relies heavily on contextualized and reflective interpretation, the question becomes: How can researchers develop AI-assisted coding approaches for critical discourse analysis that maintain interpretive depth, reflexivity, and methodological rigor while leveraging computational capabilities?

Goals and Methods

Recent methodological contributions to GenAI in discourse analytical workflows focus on general-purpose tools like ChatGPT (Curry et al. 2023; Garg et al. 2024). This project builds on the use of an AI-assistant in traditional CAQDAS tools to develop a strategy for coding. Drawing on the literature about the benefits of coding for CDA (MacMillan 2005; Bennett 2015), I argue that utilizing AI tools to suggest or apply codes leads to a co-created code system that allows for capturing broader aspects of the data than a human coder alone. This approach prioritizes interpretive depth and contextual sensitivity over speed and efficiency, as the AI can assist in the meaning-making process of discourse analysis by suggesting additional layers of interpretation.

Results:
Preliminary findings suggest that co-created code systems enhance analytical transparency and allow deeper engagement with the data. AI contributions, when guided by human interaction, can enrich the category system, allowing a more layered understanding of discursive structures. This project will showcase an example of a co-created code system.

Conclusions:

This approach addresses epistemological concerns about AI's "incapability of meaning making" (Jowsey et al., 2025) by positioning meaning-making as a collaborative process where AI contributes pattern recognition while humans provide interpretation, context, and critical analysis.