The Metamorphosis of Prime Minister Edi Rama: From Hate Speech to the Use of Deepfakes

Prime Minister Edi Rama has targeted journalist Blendi Fevziu, host of one of Albania’s most-watched political talk shows, Opinion, through a series of posts on social media that used images generated by artificial intelligence (AI), known as “deepfakes.” These posts were accompanied by denigrating and delegitimizing language directed not only at Fevziu personally, but at the media sector more broadly.

Deepfakes refer to the use of AI tools to produce videos, photographs, or audio recordings that appear authentic. Such technologies can be employed for various purposes, including entertainment, humor, artistic creativity, or advertising. However, alongside these uses, the fabrication of false visual or audio content can also serve harmful ends, such as disinformation, political propaganda, fraud, and, in extreme cases, abusive or non-consensual pornography.

Prime Minister Rama’s attacks on Blendi Fevziu followed statements made by the journalist during the January 8 episode of Opinion, which focused on the emergency situation caused by flooding in several Albanian cities. During the discussion, Fevziu criticized the lack of accountability shown by public officials toward journalists, emphasizing that they almost entirely avoid press conferences and direct communication with the media.

According to Fevziu, institutions increasingly operate through pre-recorded and pre-packaged videos presented as “live” appearances, commonly referred to as “tapes,” which are then distributed to media outlets for straightforward rebroadcast. He stated that he had evidence of at least seven ministers using this method, describing the practice as arrogant and dismissive toward citizens.

Instead of addressing the substance of these claims or presenting countervailing facts, the Prime Minister responded on social media by generating deepfake images of Blendi Fevziu in various locations and by commenting on the journalist’s personal life. Rama referred to Fevziu as an “influencer with a packed calendar” who produces “slander and lies.”

The use of deepfake technology by Prime Minister Edi Rama, framed as “political satire,” sets a dangerous precedent when considered in light of existing power imbalances and the influence his public position carries. The normalization of manipulated images and their dissemination to audiences of hundreds of thousands of followers creates a serious risk: it lowers the threshold for using false images without consent and turns such practices into routine behavior.

Moreover, deploying deepfakes against critical journalists diverts attention from the core issue of accountability on matters of public interest. It shifts the debate toward personalized narratives rather than substantive discussion. In this specific case, the deepfake images mock, infantilize, and diminish the journalist’s role, thereby undermining the very function of accountability.

Deepfakes and Derogatory Language Toward the Media: A Complete Package

Alongside the generation of deepfake images, Prime Minister Rama has repeatedly targeted the media through the use of disparaging metaphors to describe Opinion and the media landscape at large. The show’s studio has been labeled a “noisy pot that must boil every night” or a “media salon of the opposition’s body hair.” These expressions are not merely stylistic flourishes, but rhetorical tools aimed at trivializing the role of the media by portraying it as a space of chaos, noise, and disinformation.

Through such language, critical journalism is framed not as a mechanism of accountability, but as a low-grade spectacle driven by personal or political interests. This discourse contributes to the erosion of public trust in the media and helps construct a narrative in which critical questioning is treated as a problem, rather than as an essential part of public officials’ duties.

This tactic is commonly referred to as scapegoating theory, whereby responsibility and blame are shifted onto an individual or group—often a more vulnerable one—instead of confronting one’s own accountability or failures.

However, the Prime Minister’s problematic language toward journalists is not an isolated incident, but a recurring pattern. Recently, investigative journalist Klodiana Lala faced a similar reaction after criticizing the lack of waste management as a political failure, highlighting the absence of functioning incinerators—projects that have been linked to corrupt schemes—rather than focusing on citizens’ behavior.

Prime Minister Rama responded to Lala on the social media platform X, dismissing her comments as “astonishing ignorance.” Continuing his delegitimization of journalists and their work, he divided them into two categories: those who do not lie or disinform, whom he described as few, and a second group that spreads disinformation and slander and is “growing like plastic,” implicitly placing journalist Klodiana Lala in the latter category.

Last year, Rama was named “Troll of the Year” by the Reporting Diversity Network, an organization that monitors hate speech. This title is awarded to individuals, groups, or media outlets in the Western Balkans that disseminate hate speech based on gender, ethnicity, or other social categories. Prime Minister Edi Rama received this designation for the hateful language he used against opposition MP Gazmend Bardhi after the latter once again raised concerns about corruption in the incinerator case.

Technology and Corruption: Broader Concerns

The integration of AI into the political sphere is not new for Albania. For the first time, the country has appointed a Minister of State for Artificial Intelligence, tasked not only with digitalization of services but also with public procurement and the signing of contracts with private companies. In various public statements, Prime Minister Rama has described the AI Minister, Diella, as a mechanism that would help combat corruption and facilitate digitalization.

Nevertheless, despite these claims, detailed information on how Diella will function in practice and how it is expected to concretely contribute to anti-corruption efforts remains lacking. What is known is that the President of the Republic, after decreeing Edi Rama as Prime Minister, also delegated to him responsibility for the “establishment and functioning of the Virtual Minister of Artificial Intelligence ‘Diella.’”

The fight against corruption remains one of Albania’s central challenges. Data from international organizations that monitor corruption worldwide, such as Transparency International, rank Albania at 42 out of 100 points, one of the lowest scores in Europe. Corruption is particularly pervasive in public procurement, widely regarded as the most corrupt sector of the public administration.

At the same time, Freedom House classifies Albania as a “transitional/hybrid regime,” with a score of 68 out of 100, reflecting persistent weaknesses in democratic institutions and mechanisms of accountability.

Against this backdrop of AI misuse and troubling corruption indicators, the integration of Diella does not appear to promise meaningful solutions to Albania’s structural problems of transparency and accountability, as long as legal safeguards, independent oversight mechanisms, and clarity about how this instrument will operate in practice remain absent.

Author: Entenela Ndrevataj

Photo: Canva