Social question-and-answer platforms have observed a marked downturn in user engagement following the advent of generative artificial intelligence, and new empirical research reveals that tools like ChatGPT exert both direct and moderating influences on newcomer retention. According to a study published in Humanities and Social Sciences Communications by researcher Minhyung Kang, the introduction of generative artificial intelligence has directly reduced the likelihood that newly registered users continue participating on Stack Overflow. At the same time, the research demonstrates that foundational platform feedback mechanisms, including positive peer evaluations and direct user commentary, continue to play an essential role in sustaining member commitment.
To understand how external technological innovations intersect with community engagement, the study grounded its analysis in self-determination theory and motivational affordance theory. The author focused on five distinct motivational features embedded within the question-and-answer environment: upvotes, downvotes, collaborative edits, user profile completion, and comments. Rather than viewing the platform as a closed ecosystem, the investigation introduced generative artificial intelligence as a novel external factor. This allowed the researcher to examine not only its direct impact on newcomer survival rates, but also how the presence of external conversational agents moderates the motivational effectiveness of the platform’s native features.
The empirical investigation relied on an observational dataset tracking 10,000 newly registered Stack Overflow users, comparing retention patterns across different operational eras. To evaluate the probability of continued user participation, the author applied hierarchical logistic regression modeling to the user cohort. This analytical method enabled a structured comparison between the baseline retention mechanisms operating before the public arrival of ChatGPT and the altered dynamics that emerged after generative artificial intelligence became widely available to developers and learners seeking technical solutions online.
The statistical models revealed clear direct effects across several platform engagement metrics. Receiving upvotes or comments significantly promoted ongoing participation among new users, confirming the motivational power of positive feedback and active community dialogue. Similarly, user profile completion acted as a significant positive predictor of continued engagement. In contrast, receiving downvotes significantly diminished the likelihood of new user retention, while the overarching emergence of generative artificial intelligence independently exerted a significant direct negative effect on whether newcomers remained active on the platform.
In addition to direct effects, the regression analysis uncovered significant moderating shifts in how platform features influence newcomers in the artificial intelligence era. Following the arrival of ChatGPT, the positive relationship between user profile completion and continued participation weakened noticeably. The study suggests that this diminishing return occurs because autonomy-related needs, which users previously addressed by establishing an identity and customized presence within the community platform, may now be satisfied through external artificial intelligence tools that provide immediate, unassisted answers outside the social collective.
Conversely, the motivational effects of social interactions proved resilient in the face of technological competition. The positive influence of upvotes and comments, as well as the negative impact of downvotes, remained remarkably stable before and after the introduction of ChatGPT. The author attributes this consistency to the fact that generative artificial intelligence cannot replicate genuine interpersonal recognition and human community validation. Ultimately, the study extends the motivational affordance framework by accounting for external technologies, offering practical guidance for social question-and-answer platforms seeking to retain new contributors alongside generative artificial intelligence.
Subject of Research: Effects of generative artificial intelligence on new user retention in social Q&A platforms
Article Title: Direct and moderating effects of generative AI on new user retention in social Q&A platforms: evidence from Stack Overflow
Article References: Kang, M. (2026). Direct and moderating effects of generative AI on new user retention in social Q&A platforms: evidence from Stack Overflow. Humanities and Social Sciences Communications. https://doi.org/10.1057/s41599-026-08993-0
Image Credits: AI Generated
DOI: 10.1057/s41599-026-08993-0
Keywords: generative AI, Stack Overflow, social Q&A platforms, user retention, ChatGPT, self-determination theory, motivational affordance theory, hierarchical logistic regression, upvotes, downvotes, collaborative edits, profile completion, user comments, social interaction, external technological factors, moderating effects, community engagement
Cite Scienmag News
Courtney Benton. (November 16, 2026). Generative AI Directly Reduces New User Retention on Stack Overflow. Scienmag. https://scienmag.com/generative-ai-directly-reduces-new-user-retention-on-stack-overflow/
Courtney Benton. "Generative AI Directly Reduces New User Retention on Stack Overflow." Scienmag, 16 November 2026, https://scienmag.com/generative-ai-directly-reduces-new-user-retention-on-stack-overflow/. Accessed 16 November 2026.
Courtney Benton. "Generative AI Directly Reduces New User Retention on Stack Overflow." Scienmag. November 16, 2026. https://scienmag.com/generative-ai-directly-reduces-new-user-retention-on-stack-overflow/








