Enterprise AI Washing Detection Method Based on Dynamic Attention Mechanism

DOI

10.2991/978-94-6239-604-3_54How to use a DOI?

Keywords

AI washing; multi-source data fusion; dynamic attention mechanism; spatiotemporal analysis; enterprise behavior detection

Abstract

With the deep penetration of artificial intelligence (AI) technology across various industries, enterprises’ promotion of their AI capabilities has increasingly become a crucial means to enhance market competitiveness. However, some enterprises engage in exaggerated or even false propaganda in their statements about AI capabilities, a phenomenon academically termed “AI washing.“ Such misleading claims not only distort the market’s judgment of enterprises’ true technical capabilities but also may exert negative impacts on investment decisions, industry development, and public trust. To address this issue, this study constructs a detection framework based on multi-source data fusion and a dynamic attention mechanism. The framework integrates data from three dimensions: enterprises’ textual disclosures, temporal behavioral characteristics, and relational network information. Specifically, the textual dimension analyzes linguistic features in corporate press releases, financial reports, and patent literatures; the temporal dimension tracks the release patterns of AI-related announcements and their market responses; and the relational dimension reveals enterprises’ actual technical connections through supply chains and collaborative networks. The hierarchical feature extraction architecture designed in this study can adaptively adjust the weights of different data sources, thereby identifying whitewashing behaviors more accurately. Empirical analysis based on 856 AI-related listed enterprises shows that the proposed method achieves an accuracy of 89.3% and an F1-score of 0.872, demonstrating significant advantages over existing baseline methods. This research provides technical support for capital market regulation, investment risk assessment, and industry integrity construction.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Ying Tan
AU  - Zhanjie Wen
AU  - Haoxuan Ouyang
PY  - 2026
DA  - 2026/02/26
TI  - Enterprise AI Washing Detection Method Based on Dynamic Attention Mechanism
BT  - Proceedings of the 5th International Conference on Economic Development and Business Culture (ICEDBC 2025)
PB  - Atlantis Press
SP  - 526
EP  - 534
SN  - 2352-5428
UR  - https://doi.org/10.2991/978-94-6239-604-3_54
DO  - 10.2991/978-94-6239-604-3_54
ID  - Tan2026
ER  -