Advances in digital technology mean that the creation of visually compelling photographic fakes is growing at an incredible speed. The prevalence of manipulated photos in our everyday lives invites an important, yet largely unanswered, question: Can people detect photo forgeries?
Dr Kim Wade has been working with Dr Sophie Nightingale and Professor Hany Farid (School of Information, UC Berkeley) to investigate the extent to which people can distinguish between genuine and manipulated digital images.
In 2017, they published two studies in which approximately 1,400 people were asked to detect and locate manipulations in images of real-world scenes.
Respondents demonstrated a limited ability to distinguish between genuine and manipulated images.
Image manipulation is becoming increasingly widespread
When respondents were able to detect manipulations, they were often unable to locate the manipulation.
The findings revealed that people’s ability to detect manipulated images was positively correlated with the extent of disruption to the underlying structure of the pixels in the photo.
Somewhat surprisingly, there was no evidence to suggest that individual factors, such as having an interest in photography or beliefs about the extent of image manipulation in society, are associated with improved ability to detect manipulations. In a follow-up study published in 2019, the researchers showed that people fail to make use of reliable cues to manipulations (such as inconsistencies in shadows and reflections) to determine the authenticity of online images.
Ongoing research is exploring how best to train people to reliably recognise when digital images have been manipulated.
The results have important implications for police professionals and others working with digital evidence in legal contexts.