Threatened by AI: Analyzing User Demand on Community-Based Question Answering Platforms after the Launch of Generative AI
Published:
Recommended citation: Jingmei Zhou, Xiang (Charlie) Cheng, Bingjie Qian, Yulin Fang, 2023, Threatened by AI: Analyzing User Demand on Community-Based Question Answering Platforms after the Launch of Generative AI, working paper.
Abstract: Generative artificial intelligence (GAI) is increasingly competing with community-based question answering (CQA) platforms due to its proficiency in solving specialized problems. Empirical evidence is lacking, however, on how generative AI influences user demand in CQA platforms. Motivated by generative AI’s advantages and limitations, we examine its impact on user demand in CQA platforms in terms of total numbers of questions and number of questions across various novelty levels. We conduct a DID analysis leveraging ChatGPT’s launch and compare changes in user demand on a prominent international CQA platform, Quora, whose users have easy access to ChatGPT, and a dominant CQA platform in China, Zhihu, whose users are banned from using ChatGPT. We find that ChatGPT reduces the quantity of question posts, yet this substitution effect is absent for novel questions. Theoretical and practical contributions are discussed regarding generative AI’s differential impacts and developmental tactics of CQA platforms in response to AI’s increasing popularity.
Keyword: Generative AI, CQA platform, Novelty, Search costs