Abstract
This study implements and empirically evaluates the proposition-to-requirement mapping approach proposed in Wu, Li, & Cui (2026). Greenwashing detection in ESG reporting remains difficult because disclosures are narrative, heterogeneous, and often lack transparent evaluation criteria. This study proposes a proposition-level, framework-aligned method for analyzing ESG reports under the HKEX ESG Reporting Guide. The framework converts reports into structured text, decomposes them into atomic propositions, classifies propositions as disclosure, explanation, or residual, and semantically matches them to 192 binary disclosure requirements. Using Tencent’s ESG reports from 2022 to 2024 as a case, the results show that the proposed indicators are directionally consistent with changes in MSCI ESG ratings. The study provides an interpretable and traceable approach to ESG disclosure assessment.
Recommended Citation
Li, Qingan; Wu, Tianxiao; and Cui, Xiling, "A Large Language Model-Based Implementation of the Proposition-to-Requirement Mapping Framework for ESG Greenwashing Detection" (2026). ICEB 2026 Proceedings (Beijing, China). 1.
https://aisel.aisnet.org/iceb2026/1