Organizational Attention Structures and Corporate Employment in the Age of Generative AI: Evidence from Chinese Listed Firms

Abstract

Abstract Facing the rapid proliferation of generative artificial intelligence (generative AI), why do some firms intensify layoffs of administrative staff while others achieve overall workforce expansion? Grounded in the attention-based view (ABV), we investigate whether a firm’s pre-shock resource allocation channels managerial attention, thereby shaping corporate employment decisions following the commercial arrival of generative AI. Exploiting the late-2022 release of ChatGPT as an exogenous technological shock and utilizing panel data from Chinese CSI 800 listed firms from 2020 to 2025, we construct a factorial difference-in-differences (FDID) framework. Empirical results indicate that firms with an attention structure tilted toward R&D innovation prior to the shock exhibited greater employment resilience and net workforce growth, an advantage that widened steadily over time. In contrast, results for baseline absolute R&D intensity and the absolute SG&A expense ratio are statistically insignificant. Mechanism tests suggest that this innovation-oriented attention structure creates additional labor demand downstream that more than offsets the automated displacement of clerical back-office personnel. By challenging deterministic employment narratives, this study suggests that for policymakers concerned with employment insecurity, designing institutional incentives that encourage firms to focus on exploratory innovation rather than short-term cost-cutting is far more effective than relying solely on individual-level worker reskilling.