Deep Fakes: The dark side of artificial intelligence

New technologies and digital media have become an integral part of everyday life. But the combination of the two is currently giving rise to a new type of threat known as "deep fakes." More and more frequently, extremely realistic-looking fakes of photos and videos generated with the help of artificial intelligence are appearing on the Internet and social media. What brings creative potential for the film and entertainment industry leads to completely new risks in other areas: Because "deep fakes" can systematically undermine the concept of reality and be used for disinformation and targeted manipulation of public opinion. "In our multimedia-networked world, 'deep fake' technology has enormous destructive potential and, especially in liberal democracies, has a dangerous political and social dimension that should not be underestimated," says Dr. Heinz-Werner Rapp, founder and head of the FERI Cognitive Finance Institute. The background and dangers of media manipulation through artificial intelligence techniques are described in the current Cognitive Briefing of the FERI Cognitive Finance Institute "Challenge 'Deep Fake' - How digital fakes destroy reality".

Intentional falsification of reality

Understanding what is "real" presupposes that there is objective evidence that can be used to identify and verify information beyond doubt, the analysis states. The criterion of visibility plays a special role here: what is seen with one's own eyes is considered "true. "Deep fakes" abuse this "visual credibility" by creating their own deceptively real-looking "virtual reality" that is perceived as authentic and disseminated further. "Examples of this include video recordings of politicians or celebrities whose expressions and facial expressions are digitally reshaped in such a way that the original meaning of the messages is distorted or even reversed. This opens the door to deliberate deception," says Rapp. Since "deep fakes" are hardly recognizable to normal viewers due to their technically high quality and digital forgeries are often undetectable even with the most modern means, trust in the legitimacy and authenticity of visual information would be undermined. Deep fakes" thus have a twofold effect: Fake recordings could be mistaken for genuine and - possibly even more dangerously - genuine recordings could be mistaken for fakes. "Politics and business, critical media or social groups can be deliberately discredited, manipulated and 'steered' in this way - as can the general public," warns Rapp.

Increased vigilance called for

Due to rapid progress, especially in the field of AI, the potential of "deep fakes" will increase rapidly and already bring about a completely new dimension of digital data manipulation in the near future. Another accelerating factor is the high penetration of modern societies with digital "social media", which especially allows videos an enormous - mostly even global - distribution. The combination of political ruthlessness, manipulated images and massive social media reach then creates the perfect breeding ground for a chaotic system of alternative realities that could exist in parallel and on equal footing with "true" reality. "Therein lies a very serious challenge that in the future will not only give newsrooms and investors in the capital markets sleepless nights, but could also endanger the work of governments or the cohesion of entire societies," Rapp warns. Increased transparency on the phenomenon of "deep fakes" as well as a targeted understanding of the problem at the level of politics and the public are therefore urgently needed, he says.

The GErman Cognitive Briefing of the FERI Cognitive Finance Institute on the topic "Challenge 'Deep Fake' - How digital fakes destroy reality" is available in the download area.