Deepfakes: Threats to democracy and trust in the digital age
In the context of the IQ Media initiative, project coordinator Constantinos Mourlas gave an interview to the Greek newspaper tovima.gr, discussing the lifelike videos of celebrities produced using artificial intelligence, and the potential threats that deepfakes pose to democracy and trust in institutions.
The worrying reality that our personal data can be accessed by third parties without consent and permission at any time is undeniable. The use of videos created with artificial intelligence (deepfake AI) for misleading purposes is constantly increasing. These technologies are often appearing in inappropriate contexts, such as pornographic material starring Hollywood stars. In addition, deepfakes are used in advertisements and telemarketing, where even figures such as the Primate of the Greek Church can be falsely portrayed as protagonists, as he recently appeared in an advertisement promoting, in his distinctive voice, a cure for arthritis and other diseases.
Deep fakes pose a modern global challenge, as technology has advanced to a point where it can deceive both human senses and fake content detection algorithms. Moreover, Constantinos Mourlas underscores the inherent difficulty in detecting deepfakes as he mentions “lip and voice synchronization is one of the first ways for a user to understand if the video they are watching is a product of artificial intelligence. But in many of the cases the effect is so plausible that this synchronization becomes difficult to detect”.
There have been numerous instances where public figures fall victim to manipulated photos and videos created using artificial intelligence. Scarlett Johansson addressed this issue in 2019, asserting that trying to protect yourself from the Internet and obscenity is a lost cause, as her face was used in (deepfake) sex videos. According to research cited by the Guardian, nearly 4,000 celebrities have been victims of deepfake pornograph. Five years later Scarlett Johansson is speaking out against OpenAI and its CEO saying that they used a voice similar to hers for their voice-activated AI software, despite the fact that she had rejected the proposal. Another notable case is Tom Cruise, whose deepfake video went viral on TikTok in 2021, counting over 11 million views within hours. Additionally, recently in late January 2024, numerous AI-generated photos of Taylor Swift posing provocatively flooded social media.
Deepfakes have the power to create false narratives attributed to public figures, just as happened in the case of Barack Obama “admitting he is an alien”. In the form of political propaganda, they significantly facilitate the manipulation of public opinion and, consequently, electoral outcomes. One of Russia’s major channels, NTV, broadcast a fake video of a top Ukrainian official in an attempt to blame Ukraine for the Crocus City Hall massacre in Moscow in March 2024. Volodymyr Zelensky was also targeted in 2022 with a deepfake video in which he appeared to cave in to Russia’s demands and call on Ukrainians to surrender their weapons.
In conclusion, professor Mourlas argued: “We are undoubtedly in a new technological revolution that resembles the birth of the Internet. There is a need to strike a balance between the opportunities and pitfalls of technological progress and to strengthen the digital literacy of older people who use social networks as their main sources of information. This is also important for younger generations, who tend to be more cautious”.
The original article has been published in Greek on the website tovima.gr: https://www.tovima.gr/print/society/i-lernaia-ydra-lftou-deepfake/
