Artificial Intelligence vs. greenwashing: this is the project launched by Consob and the University of Trento, which have jointly developed a prototype for detecting misleading messages from green bond issuers. The goal is to develop new tools to be used in the future in financial market supervision activities to verify that the sustainability requirements claimed by corporate "green" bonds are actually met.

The project is described in the Consob Fintech Research Paper Greenwashing Alert System for EU Green Bonds: the Consob-University of Trento prototype, published today on the website www.consob.it, edited by Sandra Paterlini and Andrea Nicolodi (University of Trento) and Monica Gentile, Vincenzo Foglia Manzillo, Maria Raffaella Sancilio, and Paola Deriu (Consob).

The paper, which is in English, uses "green bonds" issued in European Union countries from 2013 to 2023 as its reference database. The prototype uses advanced technologies to disclose environmental claims within texts, analyze their tone (positive, neutral, or critical), and highlight any discrepancies between stated sustainability targets and actual performance.

The project is a major step forward in supporting supervisory activities to make document analysis faster and more effective, reducing the risk of errors or subjective interpretations. The study highlights considerable time savings: for example, compared to the approximately four hours required to analyze a sustainability report using traditional methods, the algorithm takes about ten minutes. While this prototype still needs further testing on a broader sample, it is a first step towards more transparent and responsible finance, where technology empowers sustainability.

Combating greenwashing aims to bolster investor confidence, market integrity, and support the transition to a sustainable economy. This is a key priority of Consob's Strategic Plan for 2025–2027. Within the EU, this view is fully shared by ESMA.

Press Release PDF version