Authors:
Yasin Sazid
and
Kazi Sakib
Affiliation:
Institute of Information Technology, University of Dhaka, Dhaka, Bangladesh
Keyword(s):
Dark Patterns, Data Analysis, Dataset, Bangladesh, E-Commerce, User Perception, Categorization.
Abstract:
Dark patterns are deceptive design tactics that impact user decision-making. Such tactics can harm user’s finances, privacy, and lifestyle. Researchers have explored dark patterns from different perspectives in recent times. However, most of the work have been conducted in western countries. As dark patterns involve people and their psychology, explorations in other cultural contexts may generate novel insights. We carried out such a study in the local context of Bangladesh, focusing on the prevalence and user perception of dark patterns. We first examined Bangladeshi e-commerce websites for dark patterns using a combination of automated methods such as GPT-3’s in-context learning and a novel segmentation algorithm, along with manual validation to address any limitations of the automated techniques. We also surveyed Bangladeshi university students about exposure, awareness, and concern regarding dark patterns. Based on the findings of both explorations, we ranked six dark pattern cat
egories and subsequently divided them into two novel groups with distinct traits. We also found that educational background in technology make users more aware and concerned about dark patterns. 18.3% of the websites we analyzed in this study contained dark patterns, indicating prevalence of such design practices in e-commerce websites of Bangladesh.
(More)