Unveiling the impact of artificial intelligence on corporate misconduct, the perspective of information asymmetry

Does the application of artificial intelligence technology in enterprises bring all benefits and no harm? Most current work focuses on the positive effects of AI technology usage on businesses, while largely ignoring this issue. Focusing on signal theory, we test how the adoption of artificial intelligence affects the occurrence of corporate misconduct. We argue that due to the information asymmetry between large language models and businesses, the use of artificial intelligence (AI) technology may lead to increased corporate misconduct. Using data from 4144 listed companies in China, we find evidence supporting our argument. We also analyze the impact of industry digitization, enterprise digital technology use, and executive tone on this effect, and we further distinguish the effect of this effect in different situations through additional analyses. Enterprises can utilize these findings to identify their risk points in AI technology application and develop corresponding risk management strategies accordingly.