DarkDialogs: Automated detection of 10 dark patterns on cookie dialogs

Daniel Kirkman, Kami E Vaniea, Daniel W. Woods

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract / Description of output

In theory, consent dialogs allow users to express privacy preferences regarding how a website and its partners process the user’s personal data. In reality, dialogs often employ subtle design techniques known as dark patterns that nudge users towards accepting more data processing than the user would otherwise accept. Dark patterns undermine user autonomy and can violate privacy laws. We build a system, DarkDialogs, that automatically extracts arbitrary consent dialogs from a website and detects the presence of 10 dark patterns. Evaluating DarkDialogs against a hand-labelled dataset reveals it extracts dialogs with an accuracy of 98.7% and correctly classifies 99% of the studied dark patterns. We deployed DarkDialogs on a sample of 10,992 websites, where it successfully collected 2,417 consent dialogs and found 3,744 different dark patterns automatically present on the consent dialogs. We then test whether dark pattern prevalence is associated with each of: the website’s popularity, the presence of a third-party consent management provider, and the number of ID-like cookies.
Original languageEnglish
Title of host publication2023 IEEE 8th European Symposium on Security and Privacy (EuroS&P)
Number of pages21
ISBN (Electronic)9781665465120
Publication statusPublished - 1 Jul 2023
Event8th IEEE European Symposium on Security and Privacy - TU Delft, Delft, Netherlands
Duration: 3 Jul 20237 Jul 2023
Conference number: 8


Conference8th IEEE European Symposium on Security and Privacy
Abbreviated titleEuroS&P 2023
Internet address

Keywords / Materials (for Non-textual outputs)

  • dark patterns
  • online consent
  • data protection
  • privacy
  • web measurement


Dive into the research topics of 'DarkDialogs: Automated detection of 10 dark patterns on cookie dialogs'. Together they form a unique fingerprint.

Cite this