Wikipedia Contributions in the Wake of ChatGPT

2025-03-02

AI Summary

This study investigates the heterogeneous impact of ChatGPT’s release on Wikipedia’s collaborative knowledge production. Method: Leveraging text-similarity matching, we identify articles with high overlap to ChatGPT-generated content; employing a difference-in-differences (DID) design, we analyze large-scale edit and pageview logs before and after November 2022—the month of ChatGPT’s public launch—to estimate AI’s substitution effect on human contributions. Results: Articles exhibiting both high textual similarity to ChatGPT outputs and high popularity experienced significant post-launch declines in edits and views; by contrast, low-traffic or low-similarity articles showed no such change—indicating selective user substitution toward AI-generated content and an uneven suppression of online knowledge creation. This is the first study to uncover article-level heterogeneity in LLMs’ impact on open collaborative knowledge platforms, providing causal evidence on the structural transformation of human knowledge contribution in the AI era.

Technology Category

• Humans and AI: Other Foundations of Human Computation & AI
• Knowledge Representation and Reasoning: Action, Change, and Causality
• Application Domains: Humanities & Computational Social Science

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

How has Wikipedia activity changed for articles with content similar to ChatGPT following its introduction? We estimate the impact using differences-in-differences models, with dissimilar Wikipedia articles as a baseline for comparison, to examine how changes in voluntary knowledge contributions and information-seeking behavior differ by article content. Our analysis reveals that newly created, popular articles whose content overlaps with ChatGPT 3.5 saw a greater decline in editing and viewership after the November 2022 launch of ChatGPT than dissimilar articles did. These findings indicate heterogeneous substitution effects, where users selectively engage less with existing platforms when AI provides comparable content. This points to potential uneven impacts on the future of human-driven online knowledge contributions.

Problem

• Impact of ChatGPT on Wikipedia editing and viewership
• Heterogeneous substitution effects between AI and human contributions
• Uneven impacts on future human-driven online knowledge contributions

Innovation

• Used differences-in-differences models for impact estimation
• Compared similar and dissimilar Wikipedia articles
• Analyzed editing and viewership changes post-ChatGPT