Mitigating Single-Page Dark Patterns in Subscription Cancellation Interfaces

Sammanfattning

Dark patterns are interface designs that are created with the intention to manipulate users into making decisions that primarily benefit service providers rather than the users themselves. This study focuses on dark patterns within subscription cancellation interfaces, where manipulative design strategies such as Preselection, False Hierarchy, Hidden Information and Confirmshaming are commonly used to discourage users from cancelling subscriptions. Using a Design Science methodology, mitigation strategies were derived from existing literature and applied through a GPT-5.5 Instant LLM-assisted redesign approach that transformed manipulative interfaces into more transparent and user-friendly interface versions. A user study was then conducted where participants compared original subscription cancellation interfaces with redesigned non-dark versions to evaluate user perceptions and responses. The study showed that participants generally preferred the redesigned interfaces because they were perceived as clearer, more balanced, easier to understand, and less emotionally manipulative. The results suggest that participants perceived the redesigned interfaces as more transparent and less manipulative than the original interfaces. These findings indicate that the mitigation strategies and GPT-5.5 Instant-assisted redesign approach used in this study may support future research on transparent interface design.