Greenwashing Detection in ESG Reporting: Sectoral Insights and Machine Learning Model
ESG報告におけるグリーンウォッシング検出：セクター別洞察と機械学習モデル (AI 翻訳)
Manav Gangar, Nikhil D'Souza, Rishika Kapasi, Archit Gupta, Saurabh Pandit
🤖 gxceed AI 要約
日本語
機械学習を用いたグリーンウォッシング検出フレームワークを提案。教師あり分類と教師なし異常検知のハイブリッド手法を統合し、グリーンウォッシング不一致スコア（GDS）を導入。セクター別に企業のESG開示と実績の乖離パターンを特定する。投資家や規制当局にデータ駆動型の監視ツールを提供。
English
This paper proposes a machine learning framework for detecting greenwashing by combining supervised classification and unsupervised anomaly detection. It introduces a Greenwashing Discrepancy Score (GDS) to quantify the mismatch between firms' ESG claims and actual performance. Results show sectoral variation in greenwashing patterns. The framework provides regulators, investors, and policymakers with a data-driven tool to monitor corporate ESG truthfulness, supporting SDGs 12 and 13.
Unofficial AI-generated summary based on the public title and abstract. Not an official translation.
📝 gxceed 編集解説 — Why this matters
日本のGX文脈において
日本でもSSBJ基準の導入や有価証券報告書でのESG開示が進む中、グリーンウォッシング対策は投資家対応上の喫緊の課題。本論文の機械学習フレームワークは、日本企業の開示と実績の整合性を自動評価する実務ツールとして応用可能。
In the global GX context
Globally, greenwashing undermines trust in ESG disclosures and transition finance. This paper offers a scalable, data-driven detection method that could complement regulatory frameworks like the EU's CSRD or SEC climate rules, providing a quantitative basis for enforcement and investor due diligence.
👥 読者別の含意
🔬研究者:Offers a novel hybrid ML methodology for greenwashing detection with sectoral analysis.
🏢実務担当者:Provides a practical tool (GDS) to audit ESG claims and identify discrepancy patterns in corporate reports.
🏛政策担当者:Supports data-driven regulation and monitoring of greenwashing, aligned with SDG 12 and 13 targets.
📄 Abstract（原文）
The suggested framework exhibits excellent capacity for detecting the mismatch between the environmental, social, and governance (ESG) disclosure and performance. This paper proposes a machine learning-based framework for detecting greenwashing using a hybrid combination of supervised classification and unsupervised anomaly detection approaches. There has been an increasing interest among investors and regulators in understanding the environmental, social, and governance (ESG) disclosures of firms; however, there is rising concern about the possible mismatch between firms’ claims about their actions in the environmental, social, and governance sphere and their actual performance, which raises the issue of greenwashing as opposed to malpractice. It would be pertinent to inquire how the possible greenwashing may be detected. This paper suggests that a hybrid machine learning framework integrating the two mentioned approaches may be one way of addressing the problem; it explores the potential of integrating classification techniques with anomaly detection. The study introduces a Greenwashing Discrepancy Score (GDS) - a way of putting numbers on how far companies claims about their ESG credentials might be at odds with the actual facts. Some pretty robust analysis shows that this method does indeed seem to identify patterns of mismatch between what companies say and how they’re performing - and it varies by sector. For Regulators, Investors and Policymakers, the findings give them a way to use data to keep tabs on companies and make sure they’re telling the truth about their green credentials - it’s a tool that should help policymakers and legislators tackle the problem of greenwashing and meet targets for sustainability such as those in SDG 12 (Responsible consumption and production) and SDG 13 (Climate Action)
🔗 Provenance — このレコードを発見したソース
- semanticscholar https://doi.org/10.1109/icici68867.2026.11565008first seen 2026-06-21 05:45:15
🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録（無料・3キーワードまで）。
gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。