Engineered Consent in Malaysian E-Commerce: Consumer Protection Against AI-Driven Dark Patterns

Abstract

Digital platforms increasingly obtain consumer agreement through interfaces that do more than communicate information: they organise attention, privilege particular options, and create friction around refusal or withdrawal. This article examines whether Malaysian consumer law adequately protects substantive consent when artificial intelligence amplifies dark patterns in e-commerce. Using doctrinal legal research, it analyses the Consumer Protection Act 1999, the Consumer Protection (Electronic Trade Transaction) Regulations 2024, the Electronic Commerce Act 2006, and relevant Malaysian judicial approaches to misrepresentation, consent, inequality of bargaining power, and unconscionability. The analysis is contextualised through interdisciplinary scholarship on digital market manipulation and selected European and ASEAN developments. The findings show that Malaysian law can address dark patterns involving false statements or clear non-disclosure, but is less capable of responding when factually accurate interfaces steer consumers through defaults, visual hierarchy, repeated prompts, obstructive cancellation, or personalised persuasion. Formal electronic assent is therefore capable of masking autonomy impairment, while complaint-based enforcement and economically framed remedies leave cumulative design-based harm insufficiently recognised. The article argues for a design-sensitive interpretation of unfairness and consent that assesses choice symmetry, material prominence, ease of refusal and withdrawal, and the foreseeable effect of personalised interface architecture. This recalibration would strengthen consumer autonomy without treating all persuasive design as unlawful or requiring wholesale legislative replacement.