71. Measuring Sustainability Disclosure Quality with AI-Assisted Semantic Analysis: A Study of Taiwanese Technology Firms
Invited abstract in session HE-14: Ethics and OR, Public Service, Societal Complexity, and the Common Good, stream OR and Ethics, and Societal Perspectives.
Thursday, 14:15-15:45Room: HG – Hörsaal 01
Authors (first author is the speaker)
| 1. | QIYAO HONG
|
| Department of Accounting, National Yunlin University of Science and Technology (TAIWAN) |
Abstract
This study asks whether sustainability reports are written in a clear and verifiable way. It uses a large language model to read reports from Taiwanese semiconductor firms and score each paragraph for vagueness. A paragraph is considered vague when it lacks numbers, deadlines, scope, responsible parties, or evidence. The paragraph scores are combined into a firm-level index. Firms with higher vagueness scores have lower TESG scores, even after considering report length and firm size. DJSI firms are less vague. This shows that LLMs can help detect unclear ESG disclosure.
Keywords
- Sustainable Development
- Artificial Intelligence
- Financial and Management Accounting
Status: accepted
Back to the list of papers