Desafíos actuales de la Inteligencia Artificial

84 Desafíos actuales de la Inteligencia Artificial 1. INTRODUCTION. The advent of deepfakes 1 and AI-generated media constitutes another chapter 2 in the long history of content misuse and manipulation 3 . Their existence captured the collective imagination almost instantaneously and swiftly took a sinister turn 4 . It was becoming increas- ingly evident that, alongside traditional media manipulation tools, artificial intelligence was entering the mainstream. Its broader availability promised to exacerbate the production of content further. Such production does not necessarily possess an illegal or harmful character in every instance 5 . Nevertheless, considering that the accessibility of these tools can invariably function as a double-edged sword, each significant milestone in the development and utilisa- tion of AI-generated media inherently carries the potential to broaden the operational scope of malicious actors. In other words, the dual nature of these advancements means that while they can be harnessed for innovative and beneficial purposes, they simultaneously open new avenues for misuse and exploitation by those with malicious intent. 1 On deepfakes see, among many others, WESTERLUND, Mika, “The Emergence of Deepfake Technology: A Review”, Technology Innovation Management Review, (2019) 9(11), pp. 39-52. 2 For an insightful breakdown of classic mass media manipulation frameworks see HERMAN, Edward; CHOMSKY, Noam: Manufacturing Consent - The Political Economy of the Mass Media, digital edition, Lon- don: The Bodley Head Random House, 2008. 3 For a historical perspective on media manipulation and fake news see SAFIEDDINE, Fadi: “History of Fake News”, in: Ibrahim, Yasmin; Safieddine, Fadi (eds.): Fake News in an Era of Social Media: Tracking Viral Contagion, London: Rowman&Littlefield, 2020, pp. 1-26, GORBACH, Julien, “Not Your Grandpa’s Hoax: A Comparative History of Fake News”, American Journalism, (2018) 35(2), pp. 236–249. 4 By 2017, widely regarded as the inaugural year of deepfake culture, numerous instances of deepfake pornogra- phy—typically involving AI-generated face swaps—had begun to populate various websites across the internet. For an assessment of the phenomenon see REISSMAN, Hailey, “What Is Deepfake Porn and Why Is It Thriving in the Age of AI?”, available at https://www.asc.upenn.edu/news-events/news/what-deepfake-porn-and-why- it-thriving-age-ai (last access 30.07.2024). 5 On the contrary, Generative AI output can also be deployed for beneficial purposes, see, for example, LEES, Dominic, “Deepfakes are being used for good – here’s how”, available at https://research.reading.ac.uk/re- search-blog/deepfakes-are-being-used-for-good-heres-how/ (last access 30.07.2024).