Identification of A.I. Generated Deep Fake Video by Higher Education Students

The research regarding Deepfakes has been developing at a faster pace as technology to simplify the process becomes more accessible with the use of Artificial Intelligence (A.I.). Deepfakes are a part of the Fake News area of interest and, as such, have just as much impact on the current era of the Internet as the other parts that make up Fake News. This paper presents a survey of UK University Computer Science students (n=179) and tests their ability to identify a deepfake video using their mobile phone devices. The results of the survey are able to demonstrate, with statistical significance, that educated university students in the field of Computer Science failed to identify Deepfake videos even when altered to the possibility that one of three videos is Deepfaked. In fact, while being altered, the respondents gave equal red flags to all the videos and those who indicated the correct sequence were statistically less accurate than if the guesses were made randomly. This contributes to an increasing call that educating the masses may not be enough in the fight against Fake News.