The rapid expansion of technology has, generally, been viewed as a positive development for global communication. Technological growth is expected to increase public accessibility and help citizens experience more frequent communication, reliable information, and user harmony (1). Two of these ideals, however, were put at risk near the end of 2017 when one user of the popular website Reddit published a series of pornographic videos, exploiting falsified images of celebrities. The software used to create the video was an upcoming form of computer-generated imagery (CGI). “Deepfakes”, as these doctored images were called have forced governments and citizens alike reconsider the tenets for basic democracy (2).
The computer-generated imagery used to create deepfake media is often manipulated photography and videography; which is most commonly used to impersonate people. Computer algorithms analyze a facial structure from one image and apply them to another image or video (3). Anyone with a published image of themselves on the internet could be targeted by deepfake hackers. The number of possible ways this technology could be used is immeasurable, and all possibilities bring tremendous risk. Falsified images have the power to weaken judicial institutions and evidence-based accusations. They are also capable of eliminating transparency and hindering accountability within democratic institutions. Should these institutions become undermined by widespread disinformation campaigns, democracy faces great risk (2).
There are two obvious ways that deepfake technology could be used to promote dis or misinformation, the first being impersonation. At any point in time, a skilled computer expert could create a video filled with false information, lay it over a pre-existing video of a democratically elected- or authoritarian- leader, and share it with the rest of the world. This would significantly decrease the legitimacy of all videos posted of world leaders, leaving people to wonder what can be trusted.
The second possible way for deepfake technology to be used to weaken democracy would be to blame undesirable actions or poor statements on the computer-generated images. In 2018, this issue arose, and citizens of Brazil were unsure of how to move forward (2). Joao Doria, governor of Sao Paulo in Brazil, was accused of participating in questionable sexual acts, and he was identified in the video illustrating them. When he was asked about these allegations, Governor Doria said that the video was not legitimate, and he had been the target of a deepfake attack. Analysts studied the video and have not been able to prove conclusively that it was not generated by a computer (2).
This is a frightening situation. Technology of this nature provides public officials, influential citizens, the media, and anyone else with access to the algorithm, with the means necessary to lie to the public and get away with it. An informed citizenry is key for democracy and therefore both methods mentioned previously have the potential to support authoritarian leadership in their continuous grasps for power. Leaders themselves could use this method to slander opponents, decreasing the likelihood of competitive elections and increasing their chances of staying in office.
Authoritarian regimes typically have a tight grasp on the media, but any public influencer with a desire to spread misinformation could utilize this technology. A country’s leadership could make undetectable edits to news reports and social media. A government could easily screen media outlets and make changes to their reports; all that would be required to make these deepfake videos is a computer with software that is becoming increasingly more available. These actions would tremendously decrease the legitimacy of mainstream media and any regime that acted in this manor would be adopting strong authoritarian tendencies.
Fortunately, not all is bad news, a few computer specialists will argue that deepfake technology is unlikely to develop (2). They argue that the majority of human beings are aware that this technology exists and are skeptical of imaging that does not look one-hundred percent authentic. Humans are intelligent and likely to notice these things, says Director of Dutch cyber security firm, Giorgio Patrini (2). They also claim that no matter how developed the technology might become, it will still be a modified image or video that looks unlike the original.
Despite these encouraging words, the risks of deepfake technology should not be overlooked. The future of campaigning, leadership, civic engagement, and education all rely heavily on the internet. As computer technology advances and individuals become more capable of damaging the democratic systems in place, it is crucial that no one is underestimating the risks that come with computer-generated imagery- or any other form of technology, for that matter.
It is very ominous that the experts in Brazil who analyzed the video of Joao Doria were unable to conclusively prove that it was not altered technologically, as is the point made by the Dutch security firm, that videos and images that are altered will still be accessible in their original versions distinct from their fakes. Deepfake content, then, when juxtaposed with the corresponding real footage, might allow people with particularly strong political convictions to validate whichever viewpoint they already have when they choose to trust in one version over the other. This could potentially increase polarization in society as a whole and lead to the destabilization of democracy.