Cybersecurity can be impacted by the development of technology in numerous ways, but one of the most recent and most terrifying threats is the ability to deepfake images, specifically when it related to human facial imagery being used. In the article, “Testing Human Ability to Detect ‘Deepfake’ Images of Human Faces,” the authors used an online survey to test how ability of the participants to distinguish between human faces and ones that had been deepfaked, hoping to give a consensus on how well humans can tell a real face from a fake one. The results of the study found that the participants all together barely performed better than average to tell the deepfake software apart from actual human faces. In total, the participant’s overall accuracy was around 62% but that individuals typically scored anywhere between 85% accuracy and 30% accuracy. It is no wonder that deepfakes are such a threat in cybersecurity for more reasons beside technological ones, deepfakes could be used in so many ways to cause harm, and as the results of the study show, many people are unable to accurately distinguish between falsehoods and reality. Beyond mere cybersecurity risks, if these falsehoods are seen as true it gives many societal risks and psychological risks to their usage online.

If deepfakes become more common in internet scams, then all a cyber criminal would need is to take a picture from someone’s social media account and use it as bait to possibly steal someone’s identity, money, or earn someone’s trust for something worse. Furthermore, deepfake technology could be used in cyberbullying, which would exacerbate the issue, and it could lead to an increase of distrust in our society. It’s likely, that like many other cybersecurity threats, there would be an increased victimization of marginalized groups through malicious targeting from people hoping to cause pain to others. This study highlights just how terrifying the future of technology could be if using deepfakes becomes more common and aren’t properly regulated, ethical considerations need to be made when choosing to develop deepfake technology as there will be far more than just consequences if it isn’t handled with care.
 

References
Sergi D Bray, Shane D Johnson, Bennett Kleinberg, Testing human ability to detect ‘deepfake’ images of human faces, Journal of Cybersecurity, Volume 9, Issue 1, 2023, tyad011, https://doi.org/10.1093/cybsec/tyad011