A Board Interlock Network (BN) significantly reduces the practice of “AI Wash” among Chinese A-share listed companies, according to a recent study. AI Wash involves firms overstating their artificial intelligence capabilities to meet market expectations. This exaggeration poses risks to capital market efficiency and the health of the technological ecosystem.
The study, which analyzed data from 2007 to 2022, employed a Double Machine Learning (DML) model for causal inference. Researchers found that the BN, an informal governance mechanism, effectively inhibits corporate AI Wash. This effect remained consistent through various robustness and endogeneity tests.
The inhibitory effect of the BN operates by enhancing media attention, both in terms of the number and length of reports. Increased media scrutiny amplifies potential reputational costs and exposure risks for companies engaging in opportunistic disclosure. This heightened oversight discourages firms from exaggerating their AI capabilities.
The study’s findings indicate a synergistic governance effect between internal governance mechanisms, such as interlocking directorates, and external information intermediaries like the media. This interaction shapes the corporate information environment, providing a new understanding of how BNs can mitigate speculative behavior in emerging technology sectors.
The impact of board interlocks is more pronounced in specific types of firms. Companies with greater information asymmetry, such as high-technology firms, experience a stronger inhibitory effect. Similarly, state-owned enterprises, which often have more complex governance needs, also show a greater reduction in AI Wash due to BNs.
Firms facing stronger external monitoring pressures, such as those in intensely competitive markets, also exhibit a more significant inhibitory effect from board interlocks. This suggests that the BN’s role in curbing AI Wash is amplified when other market forces are already pushing for greater transparency and accountability.
The research contributes to understanding how informal governance structures can influence corporate behavior in rapidly evolving technological fields. It offers empirical evidence for the Nature journal’s early access findings, which will undergo further editing before final publication.
Future research could explore the specific characteristics of board members within these networks that contribute most effectively to media engagement and corporate transparency. The long-term implications for investor confidence and technological development in China’s capital markets remain an area for continued observation.