LAUSR

This study investigates greenwashing in the Brazilian corporate sector by analysing the gap between companies' sustainability claims and their actual practices. We introduce a novel, AI‐powered methodology based on Decoupling Theory to detect greenwashing. Our approach triangulates three distinct data sources from 2022 to 2023: (i) internal corporate discourse (ESG reports quantified via machine learning); (ii) external market perception (Merco ESG reputation index); and (iii) actual environmental practice (objective greenhouse gas (GHG) emissions data). Our econometric analysis reveals a significant discourse‐practice gap, providing strong evidence of greenwashing. We found a substantial misalignment between companies' AI‐calculated ESG scores and their external reputation rankings, with only a small fraction of firms showing consistency. More strikingly, the analysis uncovered a paradoxical positive correlation between environmental discourse and emissions: a 1% increase in a company's AI‐calculated environmental score is associated with an approximate 5.24% increase in its GHG emissions. This suggests that environmental disclosures may serve as a tool for image management rather than reflecting tangible emission reductions. These findings empirically validate key facets of greenwashing theory within an emerging market context. The results underscore the urgent need for more robust regulation, mandatory standardized reporting, and independent verification to mitigate deceptive environmental claims and foster genuine corporate sustainability.