A Critical Discourse Analysis Diagnostic Tool for Auditing Algorithmic and Social Media Greenwashing

• Abiodun Okunola

  Ladoke Akintola University Technology

  Author

Greenwashing has migrated from static sustainability reports into the dynamic, algorithmically curated spaces of social media, where vague claims, hedged language, and strategic visual framing evade conventional audit instruments. Existing detection approaches—transformer classifiers, LLM-based multi-agent systems, and multimodal benchmarks—achieve moderate performance but lack a theoretically grounded diagnostic vocabulary capable of explaining why a claim is deceptive rather than merely classifying it as such. This study develops and validates a Critical Discourse Analysis (CDA) Diagnostic Tool for auditing algorithmic and social media greenwashing, operationalizing Faircloughian lexical and pragmatic features into a structured scoring instrument. Using a mixed-methods design that combines corpus-assisted discourse analysis of 1,200 social media sustainability posts with a prospective validation exercise involving domain experts, the tool achieved 89.4% agreement with expert adjudication and a macro-F1 of 0.87 across three greenwashing typologies: vagueness, hedging, and virtue signalling. The CDA Diagnostic Tool outperformed a transformer baseline (DistilBERT, 78% accuracy) on implicit and hedged claims while producing interpretable diagnostics that identify specific discursive mechanisms. The study contributes a replicable, theory-driven audit instrument that complements computational detection pipelines and offers practitioners a defensible evidentiary basis for greenwashing findings.

• pdf

Copyright (c) 2026 Abiodun Okunola (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.