Research On the Impact of Artificial Intelligence on Corporate Earning Management

Beijing Jiaotong University Weihai Campus, Weihai City, Shandong Province, China

^* Corresponding author: 22726056@bjtu.edu.cn

Abstract

With the widespread integration of artificial intelligence (AI) into enterprise management, its impact on corporate financial behaviour— particularly earnings management—has garnered growing scholarly interest. This study employs panel data from A-share non-financial listed firms in Shanghai and Shenzhen (2005–2024) to empirically examine how AI-related discourse in annual reports influences Real Activities Manipulation (RM) and Accrual-Based Earnings Management (AM). Utilizing text analysis and fixed-effect panel regressions, the findings reveal that higher frequency of AI-related terms corresponds to more convergent RM behaviours, suggesting that digital strategy expressions may exert normative constraints on managerial financial decisions. For AM, while the influence of AI discourse is not statistically significant, a weak negative association emerges, implying a potential inhibitory effect. Furthermore, control variables such as audit quality, cash flow status, and profitability demonstrate consistent and significant effects across both models. These results highlight the moderating role of AI language embeddedness in shaping corporate earnings management practices. The study offers theoretical contributions to the discourse on digital governance and practical implications for improving financial disclosure, enhancing regulatory foresight, and guiding AI-informed corporate strategy.