Dark patterns: reclaiming autonomy in online shopping in the age of AI?

Consumer autonomy in online shopping has become an increasingly urgent concern as e-commerce platforms deploy manipulative designs to impair free and informed decision-making, known as dark patterns. While such practices are not new, their scale, subtlety, and effectiveness have been fundamentally transformed by artificial intelligence (AI). This chapter examines how AI-driven marketing systems embed dark patterns within adaptive, data-driven optimization processes, enabling personalized, automated, and continuously refined manipulation in online shopping environments. We first develop a comprehensive typology of AI-driven dark patterns, including interface interference, nagging, forced action, obstruction, sneaking, social proof, and urgency, illustrating how each is amplified through AI-enabled personalization. We then analyze the AI models that operationalize these practices, such as reinforcement learning, clustering algorithms, natural language processing, computer vision, and generative AI, and show how granular behavioral data feeds into systems that dynamically optimize manipulation interventions at scale. To synthesize these mechanisms, the chapter introduces the AI-driven dark pattern pipeline, which clarifies how platform-level value creation is achieved alongside opaque and unintended harms to consumers, including reduced deliberation, unplanned spending, and diminished perceived choice. Finally, the chapter advances a multi-layered framework for reclaiming consumer autonomy, arguing that awareness and education alone are insufficient. Instead, effective responses require the combined deployment of structural regulation, cognitive support, and technological countermeasures that operate at the same speed and granularity as algorithmic manipulation. The chapter contributes to marketing theory by connecting AI, dark patterns, and autonomy and offers implications for research, regulation, and responsible e-commerce practice.