Shaping user perception through AI personalization and dark patterns on Meta platform

Sammanfattning

Meta Platforms increasingly employs artificial intelligence (AI) to personalized content and experiences across its platforms, such as Facebook and Instagram. While this personalization can make materials more relevant and engaging for users, it also raises ethical concerns when combined with manipulative interface designs, known as dark patterns. These designs subtly influence users' decisions, benefiting the platform more than the user's own autonomy. This study examines how personalized nudges and cleverly designed choices work together to influence what users perceive, their independence, and how they behave. Adopting a qualitative approach within an interpretivist paradigm, the research explores how users perceive, interpret, and respond to the AI-powered personalization and nudges embedded in the platform's interface design. Primary data was gathered through semi-structured interviews with eight active Meta platform users, chosen through purposive sampling to ensure a range of ages, genders, and levels of digital literacy. These conversations explored participants' awareness of personalized nudges, their emotional and cognitive responses, and the strategies, if any, they employed to resist or adapt to what they perceived as manipulation. Thematic analysis was then employed to identify recurring patterns, with the findings interpreted through the lens of Nudge Theory and ongoing discussions about the spectrum ranging from beneficial nudges to detrimental "sludge." In essence, people felt a strange pull between the convenience and the “sneaky” way their actions were being guided. While they often noticed when personalization felt a bit too accurate or when the design seemed to limit their options, they usually didn't stop using the platform. Most people just accepted it as a fair trade-off for the convenience, connection, and fun. This acceptance was strengthened by features that made these influential nudges part of their everyday routines and removed any hurdles, creating an environment where their behaviour was constantly being shaped without their awareness. This thesis contributes to the literature on AI ethics, persuasive design, and user autonomy by extending Nudge Theory into adaptive, data-driven contexts. It provides a user-centred perspective on how personalization and manipulative design function as mutually reinforcing mechanisms, offering practical implications for platform designers, policymakers, and regulators seeking to balance innovation with ethical responsibility in AI-mediated environments.