Generative AI and Human Bias: Insights from Editing Behaviors on a Knowledge-Sharing Platform

The prevalent adoption of generative artificial intelligence (AI) has brought transformative changes to user behaviors, particular in knowledge-sharing context where AI can make significant role in improving content quality. However, the potential dark sides of its adoption remain underexplored. This research investigates how generative AI impacts users’ editing behavior on a knowledge-sharing platform, with a focus on its influence on users perceptions of content quality and their subsequent editing actions. Grounded in the theoretical perspective of illusory superiority, we argue that users perceive themselves as having superior capabilities in adopting generative AI and view the content they generate as superior to that of others. We collect editing records from 1466 users on Stack Exchange between September 5th, 2022, and January 29th, 2023, and employ a difference-in-differences approach to examine the hypotheses. Our findings show that users exhibit illusory superiority in their perceived proficiency with generative AI tools and the quality of AI-generated content. However, this perception diminishes when users encounter more difficult content. Moreover, generative AI leads to an increase in edits to others’ posts, especially in formatting, while reducing edits to users’ own posts. This research contributes to the literature by highlighting how generative AI reshapes individual editing behavior and extends the concept of illusory superiority into the domain of generative AI adoption. © 2025, Academy of Management. All rights reserved.