Technology shock of ChatGPT, social attention and firm value: Evidence from China

Abstract

The release of ChatGPT has attracted widespread attention and triggered fluctuations in the capital market. This study employs difference-in-differences (DID) and event study (ES) to investigate the effects of ChatGPT's release on the cumulative abnormal return (CAR) of listed companies in China. The results reveal that a series of ChatGPT launch events, including GPT-3.5 and GPT-4, have a significantly positive impact on the firm value of the companies focused on ChatGPT, with dynamic effects. In the initial two months after the release of ChatGPT, the Chinese stock market exhibited an undervaluation of GPT-focused companies, indicating information asymmetry and competitive substitution effect. With the widespread promotion of generative AI, social recognition of ChatGPT's potential value increased. This study verifies the moderation effect of social attention in strengthening ChatGPT's impact, demonstrating that a higher search index for ChatGPT enhances stock returns for GPT-focused companies. Heterogeneity tests reveal that the impact of ChatGPT is significantly positive for large or non-state-owned companies, while small or state-owned companies show no significant effect. From the perspective of labor structure, companies dominated by technical and production personnel experience positive effects, whereas those dominated by sales personnel do not. In the eastern regions with more favorable digital economic innovation environments, companies experience a notably positive impact. This paper provides a theoretical explanation and empirical evidence for the microeconomic impact of generative AI in the Chinese context, offering valuable insights for both government and firms.