BALANCING COMPLIANCE, CREATIVITY, AND ENGAGEMENT: EVALUATING THE IMPACT OF NSFW GOVERNANCE ON USER BEHAVIORS FOR GEN AI PLATFORMS

Generative AI (Gen AI) platforms enable unprecedented creative output but also introduce new governance challenges, particularly in moderating Not Safe for Work (NSFW) content. This study examines how stricter NSFW governance affects user behavior and platform dynamics, using a natural experiment on a large text-to-image platform that implemented an external NSFW filter. Drawing on technology affordance theory (TAT), we conceptualize the stricter NSFW governance as a contraction of users’ perceived action possibilities and investigate how users respond across three dimensions: subscription behavior, compliance adaptation, and content diversity. Our results show that stricter NSFW governance led to a 14% relative decline in subscription rates among users in the treatment group, reflecting economic disengagement due to a loss of affordance. Among those who remained, users reduced NSFW content and aligned with content policies, though content diversity was not significantly impaired. Furthermore, we find that NSFW attempts surged immediately after content blocking but declined over time, suggesting a phased process of affordance actualization. Finally, compliance varied across contexts, with mobile users showing weaker adaptation than PC users, and linked accounts exhibiting stronger behavioral adjustment. Theoretically, this work extends TAT to co-creative systems and demonstrates the temporal unfolding of governance-driven boundary internalization. Practically, the findings offer actionable insights for designing governance strategies that balance platform safety, user creativity, and business sustainability.