Controversies ESG integration

Greenwashing, Sustainability Reporting, and Artificial Intelligence: A Systematic Literature Review

Can AI help investors detect greenwashing in sustainability reports?

Wayne Moodaley and Arnesh Telukdarie examine the intersection of artificial intelligence, greenwashing, and sustainability research in their paper "Greenwashing, Sustainability Reporting, and Artificial Intelligence: A Systematic Literature Review".

Using bibliometric and thematic analyses of the Scopus database, they systematically map academic publications across the binary and triple intersections of these fields from 2003 to 2022.

Their main conclusions include:

  • The application of AI to greenwashing detection is strikingly underexplored, with only 16 publications at that intersection since 2016, compared to 160 for sustainability reporting and AI over the same period.
  • Over half of the sustainability reporting and AI corpus (51%) already applies machine learning tools, primarily natural language processing, to analyse corporate disclosures at scale.
  • The intersection of all three fields returned only two publications, revealing a near-total gap in research directly applicable to identifying misleading claims in sustainability reporting.
  • Greenwashing remains pervasive despite growing awareness, with a European Commission screening finding that 42% of online green claims were potentially exaggerated, false, or deceptive.
  • Academic output on greenwashing and AI is growing at 38% annually, the fastest rate among all intersections studied, signalling rising scholarly interest in technology-driven detection methods.
  • Natural language processing, text mining, and topic modelling are the most frequently used AI methods in sustainability reporting research, which begins shaping a toolkit for greenwashing detection.

This article highlights an untapped opportunity to deploy NLP tools already proven in sustainability reporting research toward systematic greenwashing detection in corporate disclosures.

The study however only covers academic literature, so practice-based AI tools and commercial solutions may already be more advanced, especially considering the increasing adoption of Ai tools since 2022.