Fidesz and AI – bringing the worst out in each other

"The election shouldn’t be about who has more computer power and who can manufacture more fake stories. [...] Every election is a contest between at least two sides, and both have the means to manipulate things in this way. It’s more a matter of ethics whether they choose to do it or not," László Palkovics, the Commissioner for Artificial Intelligence in the Hungarian government recently said in a speech, adding that he would like to see a code of ethics and a bill restricting the use of artificial intelligence for political purposes before the elections (set to be held in April 2026 – TN).

A week later, an administrator posted a video created with artificial intelligence in the Facebook group called Harcosok Klubja (Warriors' Club), which was founded by Prime Minister Viktor Orbán himself, in which Ukrainian soldiers are shown dragging a man away from his family and beating him to death. "We are glad that this video was made. A Hungarian man was killed in Ukraine during forced conscription. At the very least, Hungary owes it to József Sebestyén to commemorate him," the Government Information Center (KTK) wrote to Telex, upon our inquiry about the video.

By this point, we had already seen several disturbing AI-generated videos presented by various Fidesz propaganda outlets, and even Fidesz politicians themselves:

• In the Warriors' Club, a video was posted with Viktor Orbán's signature, in which the LGBTQ lobby is represented by gruesome black slime, spreading and infecting terrified children, who then – along with their parents – find refuge under an orange umbrella (orange is the colour of Fidesz's party logo – TN).
• Prior to this one, there was also another AI-generated video posted in the same group, in which the Prime Minister could be heard speaking while images of Hungarian soldiers – either dead or with bleeding heads – were shown on the screen.
• Máté Kocsis (the leader of Fidesz's parliamentary group) posted a video in which Péter Magyar is shown laughing loudly while cutting through the fence that runs along the Serbian-Hungarian border with a pair of pliers, and in another video, Magyar is looking at himself in a pocket mirror while Budapest's Liszt Ferenc airport collapses behind him.
• Nemzeti Ellenállás Mozgalom (en: the Movement of National Resistance,) an organization disseminating Fidesz propaganda, purchased tens of millions of views for their AI-generated videos through Facebook ads. These videos either show Péter Magyar appearing as Ken, or waving a Pride flag in the company of opposition politicians Klára Dobrev and Gergely Karácsony, or he is shown sailing on the Tisza River in a plastic bathtub with Ukrainian President Zelensky. Where this organization got the hundreds of millions of forints needed to pay for the placement of these ads is a mystery.
• But Alexandra Szentkirályi (1, 2) and the Prime Minister's political advisor Balázs Orbán (1, 2) are also fond of posting AI-generated photos, usually also featuring Péter Magyar.

And to bring things full circle, the Fidesz faction's spokesperson, Balázs Németh, speaking in his new morning talkshow called “Warriors' Hour”, recently expressed his outrage over the internet being flooded with AI-generated political propaganda videos, which he believes are life-threatening, will lead to total chaos, which is why something must be done about it.

"Is this what the campaign will be about until the elections next spring? Will it really be about who can influence public opinion with better, more brutal, more believable, more realistic AI-generated videos?" Balázs Németh wondered. He was probably not referring to Fidesz's AI invasion, but rather to satirical videos mocking Fidesz, such as those on the Éájország YouTube channel. Of course, these are not backed by advertising campaigns costing a hundred million forints and top Fidesz politicians and influencers.

It would appear that we are dealing with a classic case of preaching one thing but practicing another. But why is it such a big deal if a political force is using artificial intelligence for this purpose?

Deepfakes and safeguards

It was sometime around 2018 that the world started to worry that deepfake videos would take fake news, misinformation, and manipulation to a whole new level, and that the possibility of using artificial intelligence to create realistic-looking fake videos would effectively wipe out reality. After all, once it became possible to convert anyone's actions or words into a video, it became possible to manufacture "evidence." and conversely, if any genuine recording could be dismissed as AI-generated fake news, then truth would no longer exist, reality would no longer exist, there would only be volume, and whoever is loudest when expressing their opinion would end up “being right”. This would also spell the end of common sense, debate and democracy.

Seven years have passed since then, and things have not turned out this way. Not because of a lack of technology: in fact, technology has advanced to the point where professionals (in the fashion industry, for example) are now able to create perfect, even overly perfect illusions, while anyone else can experiment with image and video-generating AI systems for free or for next to nothing.

In fact, it was the cynical pursuit of cost effectiveness that saved the world from the apocalyptic visions predicted in 2018. This is not only true for fake videos, but for just about everything: as we get closer to perfection, the steps we can take are becoming smaller and smaller, and these smaller steps require a disproportionate amount of effort. If I can create a fake video that can fool ten percent of the population with an investment of X amount of energy, time, and money, and a hundred times greater investment allows me to create a fake video that can fool a quarter of the population, while a thousand times greater investment helps me create a fake video that can fool half of the population, it is quite likely that I will choose the first option and make a lot of them.

This is where we are now, which is why most AI-generated content is so lame, and why it's still possible to build a business on the fact that there are people who believe that a 15-story statue of Jesus made out of zucchini and plastic bottles is real, (or what's perhaps worse, they don't even care whether it is) and they are clicking “like” and commenting "amen" without a second thought.

And this brings us back to Fidesz

The government and those in its orbit first began to dabble in this kind of content back in 2023, back when AI-generated images were still funny, harmless, and meaningless, and they only kicked things into high gear this spring. This came immediately after Donald Trump discovered these capabilities of artificial intelligence for himself and, in his characteristically subdued and tasteful manner, posted an AI-generated image of himself as the Pope, as well as a video of dollar bills raining down on the Gaza Strip with a golden Trump statue in the background. We cannot say whether Fidesz was following the American example or whether the timing was purely coincidental, but it would not be unprecedented for them to copy Trump.

Fidesz's sweeping AI campaign is particularly bizarre because it is being conducted while the party is accusing its political opponents and the press of spreading fake news and engaging in aggression online. "The culture of online degradation and hatred has now come to an end," Orbán proudly declared in reference to the "Digital Civic Circles" he launched, according to him, so that those who are not cut out for constant fighting and are fed up with conflict and want peace would have an opportunity to participate in a kind of “digital conquest of their homeland”. It was around the same exact time that the AI-generated videos about Ukrainians beating Hungarians to death, gay people vomiting black slime, and similar things started being disseminated.

As an amusing side note and a similar contradiction, Orbán is constantly communicating that these Digital Civic Circles are an unprecedented success. When he reacted to 50,000 members joining in five days, he said that "there has never before been such a force mobilized in such a short time on social media," while at the same time, the post of the hugely popular Hungarian rapper, Majka, in which he harshly criticised the government, received over 220,000 likes.

As part of Fidesz's usual distortion of reality, their main argument in connection with AI is what Government Commissioner Palkovics specifically stated at Tusványos: "But everyone can see that these are AI-generated". Which is certainly not true, as evidenced by the comments under such content. In fact, Fidesz's messaging is deliberately targeting the audience that does not necessarily notice this or simply does not care: their own hardcore fan base made up of pensioners who communicate using emojis.

The reasoning is quite similar to what Megafon-influencers usually come up with when they are caught using doctored images: “but it's just a meme, that's what the internet is like, people are just joking around”. It is precisely this kind of content which normalizes digital aggression and is fueling the "culture of online degradation and hatred" over which Orbán has recently proclaimed victory.

It is a common human trait that when we want to believe a piece of news because it fits our worldview or aligns with our prejudices, we tend to not question its authenticity, even if something obviously isn't adding up. We have all seen similar situations: someone comments on a quote or story attributed to a famous person on social media, pointing out that it is fake, only to be told off with the argument that "it doesn't matter whether it's true or not, what matters is that it's such a nice thought."

A somewhat similar phenomenon is at work when it comes to AI-generated content. Let's take, for example, the video of Péter Magyar cutting through the border fence: those who created it will be satisfied if someone believes it is real; they will also be satisfied if all that reaches people through word of mouth is that "Péter Magyar cut through the fence, and there's even a video about it on the internet"; but it is just as good for them if all it does is plant or reinforce the idea in people's minds that "this is something he would certainly do in real life." It's a bit like the giant billboards showing the leaders of the opposition that flooded Hungary in the 2018 campaign with the slogan "they would tear down the border fence together," only the AI-generated video would be a whole lot more effective.

Owning the libs

A bonus twist in the whole AI-image-generation story is that, after the initial enthusiasm, using it for commercial purposes and replacing human talent and creativity with it quickly became synonymous with being cheap and tacky. It is no coincidence that AI-generated videos as a format were ultimately embraced by the meme and troll culture on the internet, where the lack of sophistication is part of the humor. This is why the most successful AI-generated content from Fidesz is content that does not take itself seriously, but instead seeks to ridicule Péter Magyar.

A major turning point in public opinion toward AI came in March of this year with the Studio Ghibli trend, when for a while, everyone was having AI-generated images drawn in the style of the legendary Japanese animation studio. Eventually, the general consensus settled on this being a mockery and a violation of art. More precisely, it did not become the general consensus, but instead marked another fault line in the culture war: on one side are those who believe that AI art has no soul and is empty, while on the other are those who couldn't care less and actually like it. Needless to say, the two sides mutually and deeply despise one another, which is why it makes sense that Donald Trump and Elon Musk would enthusiastically post AI-generated images and videos.

In their world, if you provoke outrage, contempt, or even disgust in others counts as victory, and indeed true victory, because you have "owned the libs" and "trolled the liberals." In the Hungarian context, Orbán's reaction to the mixed reception (and that's putting it mildly) of Harcosok órája (The Warriors' Hour), is a good example of this: "The opposition is in turmoil, they can't stop themselves being outraged, they are constantly criticizing and nitpicking. There can only be one reason for this: the truth has hit home." According to this logic, Orbán can only be right, because if someone says he is wrong, they are only proving how right he is.

This is the exact system of logic, the exact world that Fidesz is working to construct through the industrial-scale dissemination of AI-generated content, even if it is of extremely poor quality. In this world, the truthfulness and validity of statements is not determined by their content, but by who said them. The boundaries of reality are becoming blurred; it has now become possible to be pro-peace while fighting a war, to be a Christian who hates others, a supporter of family values who raps about raping minors, or someone who speaks out against AI-generated content while using it on an industrial scale. And, of course, one must not get hung up on contradictions and ask questions, but should instead vote for security and stability in these dangerous and chaotic times we are living in.