Dark Patterns

A Dark Pattern (DP) is an interface maliciously crafted to deceive users into performing actions they did not mean to do. DPs make customers/users unhappy and cause them to lose (digital) trust in a business. 

As the research shows, over 95% of investigated 250 popular apps contain dark pattern. On average, each app contains at least 7 dark patterns. Another research also indicates that 11% out of 11K shopping websites have the problem of dark patterns. 

As an example of dark pattern, we should all use apps that require us to subscribe to their membership to access their service. These apps normally provide a free trial while you need to provide your bank account details to get the free trial. However, what you may ignore is that they hide a small greyed sentence on the bottom of the page, saying something like “7 days free, then $84.00/year”. It means that you will automatically opt-in after the free trial. This example contains two types of dark patterns, called forced continuity and hidden information. Forced continuity forces end-users to subscribe the membership after free trials. Hidden information patterns try to hide the critical sentence and make it hard to notice.

To mitigate such prevalent issues, we use computer vision and natural language processing techniques to do two things. First, we construct a database, extracted from crowd data, to help people understand the properties, the types of dark patterns in practice. Second, we propose techniques to automatically detect dark patterns in the wild. Our techniques can not only help the end-users to protect themselves, we can also help designers to avoid designing dark patterns, and help the regulators to implement their policies and automatically examine the potential violations. 


Please feel free to contact Jieshan Chen (Jieshan.Chen@data61.csiro.au).