Bryce L. Cooper
Old Dominion University
CYSE 201S: Cybersecurity as a Social Science
Dr. Trinity Woodbury
6th October 2026

The article “Testing human ability to detect ‘deepfake’ images of human faces” articulates a substantial issue within the field of modern cybersecurity, highlighting identifiable threats relating to the identification of deepfakes across the modern internet. The authors pursue to resolve this threat by asking: Are participants able to differentiate between deepfake and images of real people above chance levels? Do simple interventions improve participants’ deep-fake detection accuracy? Does a participant’s self-reported level of confidence in their answer align with their accuracy at detecting deepfakes? By identifying rational inconsistencies within the human psyche, the article attempts to expand current understandings of both individuals and deepfakes while providing potential strategies for mitigating this substantial threat to internet authenticity.

The pertinence of deepfake deception was commonly overlooked within cybersecurity; however, with the technology’s rapid evolution, the lingering threat deepfakes impose has caused cybersecurity specialists with a lack of understanding to approach this topic with serious intent. The article utilizes social science principles such as parsimony, objectivity, and ethnic neutrality to provide an accurate but unbiased evaluation of this cyber risk. These principles ultimately serve to enhance the analyzation process needed when collecting sufficient data within cybersecurity and underscore the need to acknowledge human error in relation to technological use. Moreover, the article utilizes various concepts examined in class such as cyber victimization, victim precipitation, human factors, and Maslow’s Hierarchy of Needs as the need to be accurate when confronted with identifying deepfakes can cause one to fall victim to the deception being presented, showing how human error and social roles within the digital environment can affect future interaction with deepfake deceptions.

The article researchers recruited two hundred and eighty participants through the Prolific online platform, who were randomly assigned to one of four experimental conditions and completed the image-judgement tasks in exchange for monetary compensation. The authors’ collection of data is through a comparative analysis, where they utilize several sections and graphs that evaluate the information gathered with precision, inclusion, and accuracy. The study signifies that individuals are better but imperfect at identifying deepfakes (Bray et al., 2026) and that the simple intervention tests did not help in their conclusion. The authors stated that human error was a major factor, as an individual’s overconfidence clouded their ability to differentiate between deepfake and real images.

Marginalized groups, specifically those who are of a different ethnic group or woman, have an increased risk of deepfake deception as they can spread false narratives and nonconsensual explicit content, ultimately impacting social standing within society. The article, however, provides significant societal contributions as it provides educational insights that explain the potential misuse of deepfakes while promoting technological solutions that showcase the output of each deepfake generation methods (Bray et al., 2026) to foster authentic online environments.

To summarize, the articles’ primary contributions expose deepfake deception through a-prior statistical power and comparative analysis. By exposing human inconsistencies because of overconfidence, this research enhances the study of social sciences within the field of cybersecurity, ultimately highlighting human error within deepfake detection and offering ways to mitigate its deception while creating safer and authentic online environments.

Reference:

Bray, S. D., Johnson, S. D., & Kleinberg, B. (2023, June 23). Testing human ability to detect ‘deepfake’ images of Human Faces . Journal Of Cybersecurity. https://academic.oup.com/cybersecurity/article/9/1/tyad011/7205694?searchresult=1 Accessed: 27th Sep. 2026