Verde ma non troppo: come algoritmi ed intelligenza artificiale possono ripristinare la trasparenza nell’ESG e smascherare forme surrettizie di greenwashing

This paper offers a multi-disciplinary analysis of greenwashing practices in the context of non-financial consolidated statements, with a particular focus on the banking sector. After outlining the EU regulatory framework on ESG disclosure, the paper introduces a computational methodology based on patternrecognition techniques to systematically detect the presence of generic or potentially misleading environmental claims in sustainability reports published by Italy’s major banking groups. The results reveal a widespreaduse of expressions lacking the verifiability and coherence required under Directive 2024/825/EU. Building onthis empirical evidence, the paper explores how algorithmic tools could support supervisory authorities andmarket operators in enhancing ESG information reliability and in preventing deceptive disclosure practices.