Detection and Analysis of Dark Patterns in Modern Web Applications

Abstract

Context: Dark patterns are user-interface design strategies deliberately engineered to coerce, mislead, or manipulate users into performing actions that serve business interests at the expense of user autonomy, privacy, and informed consent. As digital commerce and online services have proliferated, so too has the prevalence of such deceptive design choices, raising urgent ethical, legal, and technical questions. Objective: This paper presents a systematic review of dark patterns across modern web applications, proposes a multi-layered taxonomy, an automated detection framework (Dark Scan), and a severity-scoring model (DPSS). Methods: We employ a mixed-method approach combining automated DOM and behavioural analysis, expert interview synthesis, and controlled user-study evidence from published literature. Results: Dark patterns are identified in 78.4% of e-commerce platforms, 65.2% of SaaS services, and 54.7% of social-media applications. Exposure increases unintended purchase rates by 34.7% and reduces platform trust by 48.3%. Conclusion: The scale and harm of dark patterns demand coordinated responses. Our proposed Dark Scan pipeline achieves 91.4% precision and 88.7% recall, offering a practical compliance-auditing tool.