Automated Detection of Deceptive Design Patterns (Dark Patterns) In E-Commerce Websites

Online shopping has made product discovery and purchasing faster, but the same interfaces can also contain deceptive design practices known as dark patterns.These patterns can create false urgency, hide important costs, make unwanted options more prominent, encourage forced actions, or make cancellation difficult.Such practices can affect informed consumer choice and reduce trust in digital commerce.This paper presents a proposed automated detection framework for identifying dark patterns on e-commerce websites using web scraping, user-interface feature extraction, rule-based checks, and machine learning.The system accepts a website URL, collects visible text and selected interface elements, converts them into structured features, and classifies possible patterns such as fake countdown timers, hidden costs, forced enrolment, misleading choice presentation, scarcity messages, and hard-to-cancel flows.Instead of returning only a binary decision, the proposed system stores evidence and produces a pattern label, confidence score, and human-readable explanation.The paper reviews existing research on large-scale dark-pattern detection and consumer protection, describes the proposed architecture, and discusses an evaluation framework using precision, recall, F1-score, false-positive rate, and category-level performance.The work is intended as a practical research direction for analysing e-commerce interfaces, including commonly used shopping platforms, without assuming that a particular platform contains a deceptive pattern unless the system provides evidence.The proposed approach aims to support transparent shopping, research, and responsible interface design.