Abstract
The Securities and Exchange Board of India (SEBI) introduced Business Responsibility and Sustainability Reporting (BRSR) in FY 2022-23, however, empirical research related to the identification of greenwashing patterns among companies listed on Indian exchanges is yet to be available in abundance. This paper suggests a multi-layered Decision-Support System (DSS) that integrates Environmental, Social and Governance (ESG) metrics, artificial intelligence, BERT-based natural language processing (NLP), XGBoost classification, and Long Short-Term Memory (LSTM) temporal analysis, to detect greenwashing in 140 companies from NSE-500 universe for FY 2021-2024. We create a composite Greenwashing Detection Score (GDS) based on three dimensions of discrepancies: text-action discrepancy (Dₜ), numeric disclosure discrepancy (Dₙ), and temporal claim drift (Dₜₛ). Our ensemble model, based on CPCB verified environmental performance data, SEBI BRSR reporting and corporate marketing content's, is able to produce an AUC-ROC of 0.940, F1 score of 90.8% and a precision of 89.6% on the test set.

This publication has 15 references indexed in Scilit: