Deepfake technologies are based on artificial intelligence and generative neural networks, which bring us new possibilities for audiovisual creation in film, education, or artistic production, while they also represent a threat to ethics and to society itself. These videos are increasingly misused for spreading various disinformation, which leads to decreasing public trust in the media. Empirical findings confirm that most deepfake content is of a pornographic character and is created without the consent of the people involved. Such misuse results not only in the violation of human dignity and the right to privacy, but also in cyberbullying and online extortion. The political and security consequences of deepfake videos are also serious, as they can endanger democratic processes and weaken public trust in the media.

The development of digital technologies and artificial intelligence in the last decade has fundamentally changed the ways in which society receives, creates, and interprets information. Artificial intelligence has entered industry, healthcare, education, as well as media and the entertainment industry. One of its most prominent products is the phenomenon of deepfake videos, which use machine-learning methods and generative neural networks (especially GANs – Generative Adversarial Networks) to create content that is visually convincing, though in reality inauthentic.

Deepfake videos represent a new level of media manipulation because they can blur the boundary between reality and fiction. Their potential appears not only in film and artistic production, but also in medicine and education. Their negative consequences, however, clearly predominate.
According to the Deeptrace report (Ajder et al., 2019), up to 96% of all examined deepfake videos had a pornographic character, and most of them were created without the consent of the people depicted. This fact highlights severe violations of individual rights and the spread of digital violence, which has become a major area of misuse of deepfake technology. In addition, cases of political manipulation and the spread of disinformation also emerge, which can threaten the stability of democratic institutions and weaken public trust in the media.

Although most discussions about deepfake technologies focus on negative aspects, there are also areas where this technology has the potential to bring social benefit. In the film and media industry, deepfake allows creators to produce visual effects efficiently and at lower cost, as well as to reconstruct historical events and deceased figures (Kietzmann et al., 2020). In education, this technology can be used for creating personalized teaching videos or realistic language dubbing that facilitates global communication. In medicine, deepfake is experimentally used in surgical procedure simulations, training healthcare staff, and for therapeutic purposes in mental health (Gómez et al., 2022).

The aim of this article is to analyze the ethical and social aspects of deepfake videos, to present both their positive and negative uses, and to identify the risks they represent for the media environment and society. The text also strives to outline possible solutions in media literacy, social responsibility of the media, and technological measures that could contribute to preventing the spread of manipulative content.

Theory

Because deepfake videos can create extremely realistic yet false visual and sound traces, they represent a significant issue for media ethics and social responsibility, as they can be misused for manipulating public opinion and spreading false information (Husovec, 2023; Msg-life, 2025). Ethical issues arise especially when there is insufficient transparency, the absence of consent by people shown in deepfake videos, and possible abuse for political or social manipulation (Koudelka, 2024; Aleksiev, 2023).

Social responsibility in this area appears as the need for legislative regulation to protect digital identity and prevent harmful use of these technologies. The impact of deepfake videos on trust in the media and democratic institutions is fundamental because it can lead to societal polarization and the endangering of public discourse (Husovec, 2023; Koudelka, 2024). For this reason, it is important to support public education in recognizing deepfake content and to develop tools for its verification (Aleksiev, 2023; NBÚ, 2023).

Heidari (2024a) states in his study that deepfake is content generated by deep learning that appears authentic to a human observer. It is therefore a combination of the terms deep learning and fake, and refers to content generated by a deep neural network, which is a subset of machine learning. Anyone can edit video and image files. This has been possible for several years now thanks to user-friendly software suites that make it possible to edit video, audio, and images. Media manipulation has become easier because of the widespread use of smartphone applications that perform automated operations, such as lip-syncing, audio tools, and face swapping.

In a further study, Heidari (2024b) states that deepfake systems can therefore create fake images mainly by replacing scenes or images, films, and sounds, which people cannot distinguish from real ones. Different technologies now allow us to change synthetic voice, images, or video. Moreover, fraud involving videos and images is now so convincing that it is difficult to distinguish fake and authentic content with the naked eye. This can lead to various problems, from misleading public opinion to using forged evidence in court. For this reason, technologies that help us distinguish reality are important.

Hu (2023) argues in his study that the development of technologies capable of generating deepfake videos is rapidly expanding. These videos can be easily synthesized without leaving obvious traces of manipulation. Although forensic detection has achieved remarkable results in high-resolution datasets, forensic analysis of compressed videos deserves further examination. Compressed videos are common on social networks, for example videos from Instagram, WeChat, and TikTok. Therefore, the basic question becomes how to identify compressed deepfake videos.

Pennycook (2019) studies in his work how reducing the spread of disinformation, especially on social media, is a major challenge. He explores one possible approach: social media algorithms should prioritize content from news sources that users rate as reliable. For this purpose, he asks whether crowdsourced credibility ratings can effectively differentiate between more and less reliable sources. He conducted two preregistered experiments (n = 1 010 from Mechanical Turk and n = 970 from Lucid), in which individuals rated familiarity and trustworthiness of 60 news sources from three categories: (1) mainstream media, (2) hyperpartisan websites, and (3) websites producing clearly false content (“fake news”). Despite substantial partisan differences, they found that laypeople across the political spectrum rated mainstream sources as much more trustworthy than hyperpartisan or fake-news sources. Although this difference was larger among Democrats than among Republicans—mainly due to Republican distrust of mainstream sources—each mainstream source (with one exception) was rated as more trustworthy than every hyperpartisan or fake-news source in both studies when ratings of Democrats and Republicans were considered equally. In addition, politically balanced layperson ratings were strongly correlated (r = 0,90) with ratings provided by professional fact-checkers. The study demonstrates that especially among liberals, individuals with higher levels of cognitive reflection were better able to reliably distinguish between high- and low-quality information sources. It also showed that removing ratings from participants who did not know that source significantly weakened the accuracy of collective judgment. These findings suggest that algorithms prioritizing content from trustworthy media can be an effective tool in the fight against disinformation on social networks.

AlNajjar (2024) addresses how acquiring solid competencies in media ethics is one of the key outcomes of media education. In a world shaped by globalization and intercultural media relations, the fact that students possess ethical competencies relevant to their careers raises critical questions regarding the quality of ethical standards. He argues that although discussions of media ethics led to many perspectives on the scope and nature of ethical competencies in the 21st century, multiculturalism promises preparation of ethical communicators. The study is based on a survey of 32 media studies students at the American University in Sharjah and demonstrates how multicultural education rooted in a broader social context with strong cultural diversity supports hybrid views of media ethics. Most respondents stated that synthetic ethical perspectives combining local and global views on media ethics would give them suitable competencies for effectively addressing local and international issues and events using solid ethical standards.

However, the negative use of this technology has strongly overshadowed its positive potential. Already in 2019, research by Deeptrace assessed that up to 96% of deepfake videos available at that time were of pornographic content, with most created without the knowledge and consent of the depicted people (Ajder et al., 2019). This fact indicates that the primary area of deepfake misuse is not political manipulation, but rather sexualized violence perpetrated through digital media. Pornographic deepfake videos not only damage the dignity and privacy of individuals, but are often also used for blackmail and cyberbullying. This trend clearly illustrates that the biggest victims of deepfake technology are ordinary people, especially women, whose image is exploited without their consent.

Research foundations

Media ethics is grounded in basic principles such as truthfulness, accuracy, impartiality, and responsibility to the public (McQuail, 2010).

In an environment where new forms of manipulative content appear, preserving these values is key to maintaining the credibility of the media. Deepfake technologies, however, fundamentally challenge these principles because they allow the creation of audiovisual materials that can appear authentic but are actually false. This leads to a situation in which classical understanding of visual evidence becomes problematic and the public begins to ask whether it is possible to trust any media content (Chesney, Citron, 2019).

From a societal perspective, deepfake videos are a threat to democratic processes and public discourse. Political manipulation through forged speeches and statements of politicians can influence voting behavior and destroy the stability of democratic institutions (Vaccari, Chadwick, 2020). Likewise, spreading false content in crisis situations (war, pandemics) can significantly disrupt citizens’ ability to orient themselves in information and make rational decisions.

A major foundation is also the question of legislative and regulatory protection. In 2024, the European Union adopted the Artificial Intelligence Act, the first comprehensive legal framework for regulating AI technologies, including deepfake content. This document introduces an obligation to label synthetic media and supports the development of tools for their detection (European Parliament, 2024). In the Slovak context, the issue is currently addressed mainly at the level of media literacy and initiatives by independent fact-checking organizations, while the legal framework remains fragmented.

The foundations for studying the deepfake video phenomenon therefore lie in three pillars: (1) the ethical principles of media, (2) social risks that affect trust and the functioning of democracy, and (3) legislative efforts at the European and national levels, which represent the first step toward protecting society from these technologies.

Methodology

The research is presented as a theoretical-analytical study whose aim is to identify the ethical and social consequences of deepfake technologies in the media environment. Methodologically, it is based on an analysis of secondary sources, primarily scientific articles published in the Web of Science and Scopus databases, as well as expert reports of independent institutions (e.g., Data & Society Research Institute, Deeptrace Report).

The literature analysis method used allowed synthesizing knowledge from several disciplines—media studies, ethics, law, and information technologies—and thereby creating an interdisciplinary perspective on the topic. Special attention was given to studies dealing with the impact of deepfake videos on public trust in media (Vaccari, Chadwick, 2020), on democratic processes (Chesney, Citron, 2019), as well as on the spread of disinformation in online environments (Pennycook, 2019).

Secondary analysis was supplemented with practical examples, especially cases of political deepfake videos that appeared in the USA and India, and data from the Deeptrace Report (Ajder et al., 2019), which document the prevalence of pornographic content among deepfake materials. This approach made it possible to connect academic knowledge with practical implications and to prepare the ground for discussion of prevention and regulation options.

Discussion

Deepfake technologies pose significant challenges for media, society, and individuals. Although their potential in education, art, and medicine cannot be disputed, the negative consequences of their misuse now clearly predominate. The discussion of deepfake therefore needs to be conducted not only at the level of technological possibilities but also in terms of ethical and social consequences.

One key issue is maintaining trust in the media. Traditionally, a visual recording was considered a reliable witness to reality. Deepfake disrupts this certainty, leading to what is called a “crisis of trust.” If a person cannot be sure that a recording corresponds to reality, it can lead to a loss of trust not only in the media, but also to weakening of public discourse itself.

Another issue is political manipulation. Examples of forged political speeches illustrate that deepfake can fundamentally influence democratic processes. Such content can be used to polarize society, spread hate, or influence elections. Similarly dangerous is the use of deepfake during crises, for example in military conflicts or pandemics, when the public is especially sensitive to disinformation.

Particular attention is required for the ethical responsibility of the media. Journalists and media institutions face the challenge of distinguishing authentic content from manipulative content. Fact-checking organizations and journalists play a key role in this process, yet they also have limits. Responsibility therefore also lies with the technology platforms that distribute content, and these are precisely the ones that should invest in tools for detecting and labeling synthetic media.

Media literacy offers an important preventive measure. As Pennycook (2019) notes, the public’s ability to distinguish between trustworthy and untrustworthy sources can significantly contribute to limiting the spread of disinformation. Educational institutions, non-profit organizations, and the state should support programs that lead citizens toward developing critical thinking and the ability to verify information.

Technological solutions offer another part of the answer. The development of algorithms to detect deepfake content, as well as introducing digital watermarks and labeling of synthetic media, represent important steps toward prevention. The European AI Act (2024) already introduces a requirement for transparent labeling of deepfake materials, which can contribute to greater user awareness. At the same time, the need for interdisciplinary research is increasing, one that links technological innovation with legal and ethical frameworks. A challenge remains that, as deepfake becomes more sophisticated, it can be increasingly difficult to distinguish fake from authentic content, even for advanced forensic tools (Mirsky, Lee, 2021).

The discussion thus demonstrates that an effective response to the deepfake phenomenon must be interdisciplinary. It is not enough to rely solely on technical solutions; it is necessary to connect the technology sector, media, legislation, and academia. In this way, it is possible to achieve a balance between creative freedom and protecting society from manipulative content.

Conclusion

The deepfake video phenomenon represents one of the greatest ethical and social challenges of the contemporary media environment. On one hand, it is a technology that opens new opportunities in film, education, and art; on the other hand, it poses substantial risks to media credibility, the stability of democratic processes, and individual safety. Our analysis has shown that the main problem is not the mere existence of deepfake technology, but its misuse to spread disinformation, cyberbullying, or political manipulation.

The media today face the challenge of how to maintain public trust in an environment where almost any visual or audio recording can be falsified. Responsibility in this regard is borne not only by journalists, but also by the technology platforms that distribute content. Education and raising citizens’ media literacy also play a key role, as it can significantly contribute to societal resilience against manipulative content.

In conclusion, it can be stated that an effective response to the challenges associated with deepfake videos requires an interdisciplinary approach. Linking technological development, legislative measures, media ethics, and academic research represents the path toward creating a framework to protect society from the negative consequences of this technology. If it is possible to find a balance between freedom of creation and protection of the public interest, deepfake can remain a tool for creativity rather than a threat to democracy and trust in the media.

Authors:
Mgr. Dominik Maček, internal doctoral student
PaedDr. Beáta Pošteková, PhD.
Žilinská univerzita v Žiline, Ústav mediamatiky a kultúrneho dedičstva, Žilina
Affiliation: This is a partial output of the VEGA grant task 1/0324/25 Longitudinal Research on Media Literacy in the Context of Sustainable Social Development under the conditions of Slovakia.

References

AJDER, H., PATRINI, G., CAVALLI, F. a CULLEN, L., 2019. The State of Deepfakes: Landscape, Threats, and Impact. Amsterdam: Deeptrace. Deeptrace Report.

ALEKSIEV, A., 2023. Expert advises what to watch for in deepfake videos. Trend.sk [online]. Bratislava: Trend Holding. [accessed 2025-10-04]. Available at: https://www.trend.sk/technologie/vyskyt-deepfake-videi-vyrazne-rastie-odbornik-radi-co-davat-pozor-co-vsimat

ALNAJJAR, A., 2024. Gaining Solid Competence in Media Ethics: A Multicultural Perspective. Journal of Media Ethics. 39(1), 15–28. [online]. [accessed 2025-10-04]. Available at: https://www.webofscience.com/wos/woscc/full-record/WOS:001282396600001

CHESNEY, R., CITRON, D. K., 2019. Deep fakes: A looming challenge for privacy, democracy, and national security. California Law Review. 107(6), 1753–1819.

EUROPEAN PARLIAMENT, 2024. Artificial Intelligence Act. Brussels: European Union.

GÓMEZ, J., MARTÍN, M., LÓPEZ, C., 2022. Applications of Deepfake Technology in Healthcare: Opportunities and Challenges. Journal of Medical Systems. 46(7), 1–10.

HEIDARI, A., 2024a. Deepfake technology and human perception. Journal of Information Security Research. 12(2), 55–67. [online]. [accessed 2025-10-04]. Available at: https://www.webofscience.com/wos/woscc/full-record/WOS:001148512000002

HEIDARI, A., 2024b. The implications of deepfake systems for law and society. International Journal of Law and Technology. 18(1), 22–36. [online]. [accessed 2025-10-04]. Available at: https://www.webofscience.com/wos/woscc/full-record/WOS:001107488700001

HU, J., 2023. Forensic detection of compressed Deepfake videos on social networks. Multimedia Tools and Applications. 82(3), 4125–4140. [online]. [accessed 2025-10-04]. Available at: https://www.webofscience.com/wos/woscc/full-record/WOS:000766700400017

HUSOVEC, M., 2023. Deepfakes and privacy protection. [s.l.]: [s.n.].

KIETZMANN, J., LEE, L. W., MC CARTHY, I. P. and KAHN, A. M., 2020. Deepfakes: Trick or treat? Business Horizons. 63(2), 135–146.

KOUDELKA, M., 2024. Fake videos are flooding the internet. Česká televize [online]. Prague: Česká televize. [accessed 2025-10-04]. Available at: https://ct24.ceskatelevize.cz/clanek/domaci/falesna-videa-zaplavuji-internet-spolehlive-nastroje-na-rozeznani-deepfaku-ale-neexistuji-348484

MCQUAIL, D., 2010. McQuail's Mass Communication Theory. 6th ed. London: Sage. MIRSKY, Y. and LEE, W., 2021. The Creation and Detection of Deepfakes: A Survey. ACM Computing Surveys. 54(1), 1–41.

MSG-life, 2025. Deepfake significance: How to uncover and recognize it from the original [online]. Bratislava: Msg-life. [accessed 2025-10-04]. Available at: https://msg-life.sk/clanky/digitalizacia/deepfake/#:~:text=vr%C3%A1tane%20ochrany%20osobn%C3%BDch%20%C3%BAdajov

NÁRODNÝ BEZPEČNOSTNÝ ÚRAD, 2023. NBÚ warns about deepfake videos, explains what to watch for. Týždeň [online]. Bratislava: NBÚ. [accessed 2025-10-04]. Available at: https://www.tyzden.sk/politika/98941/nbu-upozornuje-na-deepfake-videa-uvadza-co-si-treba-vsimat/#:~:text=nezvy%C4%8Dajn%C3%A9%20pohyby%20tela%20a%20tv%C3%A1re

PENNYCOOK, G. a RAND, D. G., 2019. Fighting misinformation on social media using crowdsourced judgments of news source quality. Proceedings of the National Academy of Sciences. 116(7), 2521–2526.

VACCARI, C. a CHADWICK, A., 2020. Deepfakes and disinformation: Exploring the impact of synthetic political video on deception, uncertainty, and trust in news. Social Media + Society. 6(1), 1–13.