Dark Patterns Detection on E-Commerce Websites
Aarya Sawant, Amrisha Gamane, Shraddha Sonawane, Omkaresh S. Kulkarni, Rutuja Rajendra Patil, Gagandeep Kaur · 2025
Dark patterns are techniques employed on e-commerce website interfaces for the purpose of influencing users' actions to engage in undesirable behavior such as making unintentional purchases or submitting personal information when they did not want to. This is where this study seeks to address through identification of such patterns using a machine learning method. To cover false and genuine content, a dataset with 10,000 textual samples, obtained from several e-commerce sites, was compiled. Thus, for analysis, the easily interpretable classification model, Logistic Regression, was used. The results showed that the detection accuracy of the model is 92%, where precision and Recall metrics' point to the model's abilities to detect manipulative language. The results point to the ability of machine learning methods in detecting and combating fraud in cyberspace and thus promoting an ethical e-commerce market.