Abstract:

The release of ChatGPT in November 2022 constitutes a major technological shock, rapidly expanding firms’ access to large language model (LLM) capabilities. While prior research documents task-level productivity gains from generative AI, its firm-level impact on innovation remains unclear. We argue that the effects of generative AI depend not only on technological exposure but also on internal organisational conditions.

We examine whether ex ante LLM exposure influences firms’ innovation efficiency following the ChatGPT shock and whether employee-perceived organisational culture moderates this relationship. Using a panel of US research-intensive public firms (2019–2024), we measure industry-level LLM exposure using pre-shock occupation-based indices and capture innovation efficiency through Research Quotient (RQ), the firm-specific output elasticity of R&D. Organisational culture is proxied by pre-period firm-level ratings from Glassdoor. We implement Difference-in-Differences and Triple Difference specifications exploiting cross-industry exposure variation and cross-firm heterogeneity in culture.

While we find no significant difference in immediate average innovation outcomes between highly exposed and less exposed firms, substantial heterogeneity emerges across firms with different levels of employee-perceived organisational conditions. Among highly exposed firms, those with stronger employee-perceived organisational culture exhibit relative improvements in innovation efficiency, whereas firms with weaker culture experience muted or adverse adjustments.

These findings indicate that generative AI is not uniformly productivity-enhancing. Instead, its innovation impact depends on organisational complementarities. Firms with stronger internal organisational conditions appear better positioned to convert AI exposure into research efficiency gains. By identifying organisational culture as a key moderator of technological adaptation, this study advances understanding of how firms respond to disruptive AI shocks and contributes to research on innovation, strategy, and organisational capabilities.

Please choose your submission type: Developmental Paper

Primary Track: Innovation

Menu

Dr. Di Xiao