In an article for the Athens-Macedonian News Agency, the Special Secretary for Data Governance and Artificial Intelligence, Vasilis Karkatzounis, presents Greece’s contribution to the European Union’s strategy on deepfakes.
Deepfakes are one of the most characteristic yet alarming manifestations of modern Artificial Intelligence (AI) systems. They refer to audiovisual content (image, audio, or video) that is generated or modified by AI and convincingly resembles real people, events, or locations, creating a false impression of authenticity. Deepfakes can be used for satire or entertainment, but at the same time they pose serious challenges to the reliability of information, the protection of fundamental rights, and the functioning of democratic institutions.
The proliferation of deepfakes in recent years has been explosive, blurring the lines between the real and the artificial. According to research, deepfake videos increased by 550% between 2019 and 2023, while by the end of 2025, it is estimated that up to 8 million deepfakes will have been shared online, up from approximately 500,000 in 2023—a figure that nearly doubles every six months. Meanwhile, it is estimated that over half of the content on the internet is already AI-generated, with users—even experts in the field—finding it increasingly difficult to distinguish the synthetic from the real.
The social implications of deepfakes are multidimensional. First, in the realm of democracy and diplomatic balances, the ability to convincingly replicate political figures or public events can undermine public discourse, influence electoral processes, and facilitate organized disinformation campaigns, turning AI into a tool for hybrid threats. Second, in terms of security and the economy, deepfakes are increasingly being used in large-scale financial fraud, impersonation, and identity theft, making criminal activity more convincing and harder to detect. Third, regarding individual rights and privacy, a large portion of deepfakes involves non-consensual sexual content, overwhelmingly targeting women, with serious consequences for the dignity and mental health of the victims. Finally, trust in information is undermined overall: the so-called “liar’s dividend” leads citizens to question even authentic material as a “possible deepfake,” causing a widespread crisis of credibility.
Read the full article at https://www.amna.gr/home/article/972097/.
