Generative Artificial Intelligence and ESG Reporting: Efficiency Gains, Trust Risks, and a Governance Perspective
生成AIとESG報告：効率性の向上、信頼リスク、およびガバナンスの視点 (AI 翻訳)
Xinyi Wang
🤖 gxceed AI 要約
日本語
本概念論文は、ESG報告における生成AIの応用を批判的に検討する。生成AIはテキスト処理と開示文書作成のコスト・時間を削減し効率を高める一方、透明性・信頼性・説明責任の課題を生む。特に、効率と信頼のパラドックスを指摘し、AI支援による説得力のある開示がグリーンウォッシュのリスクを高めると論じる。企業・業界・規制の多層ガバナンス枠組みを提案する。
English
This conceptual paper critically examines generative AI in ESG reporting. It argues that while generative AI enhances efficiency by reducing cost and time in processing text-intensive data and drafting disclosures, it introduces challenges to transparency, reliability, and accountability. The paper identifies an efficiency-trust paradox, where AI-assisted disclosures may be more persuasive yet potentially misleading, increasing greenwashing risk. It proposes a multi-level governance framework involving firm-level practices, industry coordination, and regulatory oversight.
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
日本ではSSBJ開示が始まり、有報・統合報告書でのAI活用が注目される。本論文の効率と信頼のパラドックスは、日本企業がAIを活用した開示の信頼性を確保する上で重要な示唆を与える。また、金融庁の開示規制や監査の枠組みにAIガバナンスを組み込む議論に寄与する。
In the global GX context
Globally, with ISSB and CSRD mandating detailed sustainability disclosures, AI tools are increasingly used to streamline reporting. This paper highlights the risk of AI-generated disclosures undermining trust and increasing greenwashing, which is critical for regulators and standard-setters. The proposed governance framework offers a foundation for integrating AI oversight into global disclosure frameworks.
👥 読者別の含意
🔬研究者:AIとESG報告の交差における効率と信頼のパラドックスを理論的に整理し、今後の実証研究の基盤を提供する。
🏢実務担当者:AIを開示プロセスに導入する際のリスク管理とガバナンス設計の指針を得られる。
🏛政策担当者:AI時代の開示規制と監査の枠組みにAIガバナンスを組み込む政策的示唆を得られる。
📄 Abstract（原文）
This conceptual paper critically examines the application of generative artificial intelligence, hereafter referred to as generative AI, in environmental, social and governance (ESG) reporting. As sustainable development rises in global prominence, ESG disclosure has become an integral component of corporate reporting, prompting firms to seek more efficient ways to process and disclose large volumes of non-financial information. The analysis shows that generative AI can significantly enhance the efficiency of ESG reporting by reducing the cost and time associated with processing text-intensive data and drafting disclosures. However, these efficiency gains entails unprecedented challenges related to transparency, reliability, and accountability. In particular, the use of the opaque algorithms and input-dependent outputs raises concerns about the verifiability and credibility of ESG information. More importantly, this paper identifies an efficiency–trust paradox in AI-assisted ESG reporting. While generative AI improves reporting efficiency and narrative coherence, it may simultaneously weaken stakeholder trust by enabling more persuasive yet potentially misleading disclosures, thereby increasing the risk of greenwashing. Based on these findings, the paper develops a multi-level governance framework that encompasses firm-level practices, industry-level coordination, and regulatory oversight. This framework seeks to mitigate the risks associated with AI adoption while preserving the integrity of ESG reporting. The study contributes to a more nuanced understanding of how generative AI reshapes ESG disclosure practices, and highlights the need to balance technological efficiency with trust, authenticity and accountability.
🔗 Provenance — このレコードを発見したソース
- semanticscholar https://madison-proceedings.com/index.php/aemr/article/download/5245/5247first seen 2026-08-17 05:22:01 · last seen 2026-08-18 04:50:20
🔔 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。