Project Title: When Humans Meet Deepfakes: From the Lens of Eye-Tracking and Self-reported Data[cite: 1]

Ethics Approval ID: 2026177926531610645[cite: 1]

About the Study

Deepfake content is increasingly common online and is frequently used for misinformation, manipulation, and fraud[cite: 1]. As deepfake images grow more realistic, understanding how people tell authentic from manipulated visuals is essential[cite: 1]. This study investigates how people distinguish deepfakes from genuine images using eye-tracking and self-reports[cite: 1]. It explores whether visual patterns and reasoning influence detection, hypothesising that cues aid recognition and that brief education improves accuracy[cite: 1].

Participation Criteria

• Inclusion: Participants must be aged 18 years or older, can read and understand English, and have normal vision sufficient for viewing images and completing eye-tracking calibration[cite: 1].
• Exclusion: You cannot take part if you are under 18 years of age, cannot read or understand English, or have visual impairments that would interfere with viewing images or accurate eye-tracking measurement[cite: 1].

Data Protection & Confidentiality

Information collected during this study will be held confidentially by the researchers in line with the UK Data Protection Act 2018[cite: 1].

• Only researchers involved in the study will be authorized to access the data, but anonymised responses may be shared with other researchers or made available in online data repositories[cite: 1].
• To ensure confidentiality, your data files will not include identifying personal information (or you will use a unique participant identification number)[cite: 1]. Any identifying details will be stored separately and securely[cite: 1].
• Anonymized data will be held securely on University of Kent computers or secure partner platforms[cite: 1]. Any publication resulting from this work will report only aggregated findings or fully anonymized examples[cite: 1].
• Participation is entirely voluntary, and you are free to withdraw at any time during the survey without giving a reason by closing your browser[cite: 1].

Contact Us

If you have any questions or concerns about this study, please feel free to contact the researcher:

David Boamah — dab54@kent.ac.uk

This study consists of three core stages designed to analyze your gaze patterns and detection reasoning[cite: 1]:

• Eye-tracker Calibration: To allow accurate recording of your visual attention[cite: 1].
• Pre-training Detection Session: You will view 10 images (a mix of deepfake and non-deepfake images) and classify each as either "deepfake" or "not deepfake"[cite: 1]. You will also provide a brief written explanation justifying your decision after each image[cite: 1].
• Training Session: You will watch a short educational video highlighting common indicators of deepfake images[cite: 1].
• Post-training Detection Session: You will view a new set of 10 images and repeat the classification and explanation process[cite: 1].
• Final Questions & Debrief: A short concluding section[cite: 1].

Important Testing Guidelines:

• There is no time limit per question. Please take the time you need to examine the details carefully before making an accurate choice[cite: 1].
• Please do not refresh the page, go back, or close the browser window during the active testing sessions to avoid losing your progress.

Practice Examples

Before starting the official evaluation, you will complete a quick guided sample to get comfortable with the interface buttons and image zoom controls.

Click "Continue" to start the example session.