The Potential for Unknowingly Disclosing Personal Information via Eye Tracking Technology

Callum Woods

Research output: ThesisDoctoral Thesis

181 Downloads (Pure)


We evaluate whether personal information, such as an individual’s personality, gender,
or self-esteem can be predicted from their visual behaviour upon social networking
site (SNS) based content. This SNS context provides an ecologically valid,
and novel, visual environment and behaviour upon such sites has been found to reflect
a wide range of personal attributes. Our novel contribution to the literature is
to highlight that, through the use of machine learning techniques, visual behaviour
provides insight into a range of personality traits and personal attributes within very
short (sub minute) time scales. This is in contrast to previous approaches that aggregate
digital logs of SNS behaviour across weeks, months or years of use to make
similar predictions. Furthermore, we evaluate which types of visual behaviour are
most informative when predicting personal information and find that, in certain situations,
it appears that it is not critical to know the type of content being displayed
upon the page. We highlight that this has important implications for privacy, especially
with eye tracking becoming increasingly popular as a way for users to interact
with their computer.
Original languageEnglish
Awarding Institution
  • Royal Holloway, University of London
  • Durant, Szonya, Supervisor
  • Luo, Zhiyuan, Supervisor
  • Watling, Dawn, Supervisor
Thesis sponsors
Award date1 May 2021
Publication statusUnpublished - 2021


  • Social Media
  • Eye Tracking
  • Privacy
  • machine learning
  • Personality Assessment

Cite this