Dark Patterns Detection and Classification Using Machine Learning Techniques

UX design shapes the interaction with a digital platform. Most practices., while attempting to increase user satisfaction., can actually deceive users using so-called dark patterns in order to gain an advantage for the service provider. This paper addresses the automated classification of five dark patterns: Urgency., Obstruction., Social Proof., Misdirection., and Scarcity., along with ethical UX practices., all labeled in English. First., different machine learning algorithms will be applied., such as Logistic Regression., Support Vector Machines., and Random Forest and Gaussian Naive Bayes with Grid-search for hyper parameter tuning. Then., the discussion of accuracy., confusion matrices., and AUC-ROC scores will present how such models can work toward the detection of dark patterns in order to foster ethical design.