Paper Type
Short
Paper Number
PACIS2025-1732
Description
With rapid advancements in AI-based search, digital knowledge-sharing platforms are experiencing a profound transformation. Traditional Q&A sites like Stack Overflow are integrating AI tools to enhance search precision and streamline information retrieval, yet little is known about how these innovations affect user engagement. Motivated by this gap, we investigate the impact of Stack Overflow’s OverflowAI using a quasi-experimental design that compares the post-launch period in 2023 with a control period from 2022. Our analysis reveals that OverflowAI is associated with an approximately 18% increase in post view counts and a 3% increase in vote counts, with heterogeneous effects across technical topics. These findings underscore the potential of AI-driven search to not only improve the relevance of retrieved information but also to foster greater community interaction. Our results provide valuable insights for platform managers and developers in optimizing content curation and engagement strategies in digital knowledge ecosystems.
Recommended Citation
Park, Seokran and Kim, Dongyeon, "Impacts of AI-Based Search on User Engagement: Evidence from Stack Overflow’s OverflowAI" (2025). PACIS 2025 Proceedings. 22.
https://aisel.aisnet.org/pacis2025/aiandml/aiandml/22
Impacts of AI-Based Search on User Engagement: Evidence from Stack Overflow’s OverflowAI
With rapid advancements in AI-based search, digital knowledge-sharing platforms are experiencing a profound transformation. Traditional Q&A sites like Stack Overflow are integrating AI tools to enhance search precision and streamline information retrieval, yet little is known about how these innovations affect user engagement. Motivated by this gap, we investigate the impact of Stack Overflow’s OverflowAI using a quasi-experimental design that compares the post-launch period in 2023 with a control period from 2022. Our analysis reveals that OverflowAI is associated with an approximately 18% increase in post view counts and a 3% increase in vote counts, with heterogeneous effects across technical topics. These findings underscore the potential of AI-driven search to not only improve the relevance of retrieved information but also to foster greater community interaction. Our results provide valuable insights for platform managers and developers in optimizing content curation and engagement strategies in digital knowledge ecosystems.
Comments
AI ML