Generative AI and Human Knowledge Sharing: Evidence from A Natural Experiment

Generative AI, known for its content creation capabilities, possesses great potential to reshape knowledge-sharing activities. By leveraging a unique policy of a leading online Q&A platform that introduces generative AI answers, we explore the effects of generative AI answers on human knowledge contribution efforts in knowledge-sharing platforms. Our findings demonstrate the positive effects of the generative AI answers on human knowledge contribution efforts, both in terms of quantity and length of the subsequent answers to a question. Notably, a more nuanced investigation into human-AI answer similarity indicates that generative AI answers prompt users to provide answers more aligned with the AI-generated answers, underscoring the AI-conformity effect rather than the AI-differentiation effect as the underlying mechanism. Moreover, we observe an increase in human experts’ contribution following the introduction of generative AI answers, suggesting generative AI answers encourage rather than crowd out human experts’ contribution to the platform. In a follow-up randomized experiment, we offer corroborative evidence for the results of empirical analysis. Furthermore, it provides preliminary evidence that labeling generative AI as the knowledge provider, rather than the content of generative AI answers, primarily drives the observed effects of introducing generative AI answers. Our research highlights the potential of generative AI to motivate human participation in knowledge generation and dissemination, adding to the burgeoning body of work on how generative AI influences human knowledge sharing.