Why develop your political motion’s ideology when you may generate it immediately with synthetic intelligence? Consideration spans are quick anyway. Complicated narratives are redundant, aren’t they? Feeding off stereotypical visible references to an ideological previous is sufficient, proper?
AI-generated aesthetics have gotten inseparable from a globalized type of up to date extremist narratives leveraged by far-right political events. The wording utilized in prompts for big language fashions (LLMs) to create idealized digital photos transforms the perceived id of the nationalist ‘in-group’ – and impacts actuality. In Romania and Moldova, for instance, the dimensions of AI-bot-flooded far-right election campaigns set alarming precedents in 2025. Equally, AI enhances the at present efficient German, far-right AfD celebration’s rhetoric.
Grassroots Slovak nationalism
Small nationalist teams are equally energetic in creating their idealized id, including to the morass of AI-generated imagery to extend their supporters’ perceptions of belonging. Whereas additionally extremely generalized, their imagery faucets into localized historic references.
The Slovak, pro-Russian bikers membership Brat za Brata (Brother for Brother), for instance, commonly posts AI-generated photos on Fb resembling footage of younger, white, blond women and men wearing conventional costume, of muscular ‘medieval’ Slavic brotherhood, and of anthropomorphized wild animals representing Russia, Slovakia and Czechia standing shoulder to shoulder. The photographs are sometimes set in approximations of the good outdoor, dramatically lit for heroic impact. Up to date propaganda aesthetics are a layered jumble of components: generic depictions of nature and stylized folks meet nationalist symbolism inside fantasy settings harking back to CGI films for the entire household. Historic stereotypes don’t disappear. They’re closely referenced inside new output. AI simply makes them simpler to generate shortly and repetitively.
Photos from Brat za Brata Fb web page.
In distinction, precise footage of Brat za Brata present a gaggle of middle-aged males posing collectively in city areas, sporting black bomber jackets emblazoned in Cyrillic. The one fixed between the 2 units of images is that the bikers are flying the Russian flag in each.
Whereas claiming to be apolitical, Brat za Brata, which organizes Soviet memorialization occasions, have been utilizing their on-line platform to actively promote a pro-Russian agenda for the reason that full-scale invasion of Ukraine. The motion’s chief, Matúš Alexa has met Sergei Naryshkin, the top of the SVR, Russia’s international intelligence service. Members commonly attend occasions on the Russian embassy in Bratislava, boasting of their relationship with former ambassador Igor Bratchikov. In March 2025, the group delivered a questionable petition to the Slovak president calling for sanctions towards Russia to be withdrawn.
Picture from Brat za Brata Fb web page.
The gentle energy of a cartoon bear
Whereas saturated in seemingly healthful innocence, Brat za Brata’s representations of a ‘true’ Slavic id take their place amongst an onslaught of different AI-generated far-right imagery that advocates for strict nativism. Russian state-owned media, aficionados of AI-generated propaganda, commonly hit out on the legitimacy of worldwide democratic establishments, particularly in relation to Ukrainian membership. In a sequence of provocative photos posted by RIA Novosti, rats with demonic pink eyes and bared fangs clammer inside a field labelled ‘NATO’ whereas others sporting the Ukrainian flag comply with the pied piper or stage a peasants’ revolt beneath an arch of EU stars.
Though no single social media put up may be capable of shift public opinion, collectively they are often extremely instrumental. The actual potential of AI-generated imagery is in its openness to the incremental radicalization and polarization of id and myth-building. ‘When you don’t assemble a grand narrative – saying one thing right here and there however by no means permitting the total, harmful image to return collectively –’ it’s ‘troublesome to use the legal guidelines we generally have in Europe concerning hate speech …, as a result of all the pieces stays merely suggestive,’ says thinker of language Tomáš Koblížek.
AI-generated photos utilized by RIA Novosti.
Even less-explicitly pro-Russian-narrative content material is subsequently additionally inflicting controversy. Just lately, the extensively out there Netflix youngsters’s present Masha and the Bear, for instance, confronted a backlash in Ukraine and the UK for likening its ‘gentle-giant’ bear to Russia. Making the knowledge area friendlier to broadly pro-Russian narratives will be an train in gentle energy. In avoiding strictly political, polarizing opinions, gentle energy enhances conventional diplomacy, speaking by way of civil society channels.
Some propaganda imagery is intentional nonconfrontational, whether or not it takes the type of the bear from a child’s cartoon or Brat za Brata’s ‘Disney-inspired’ Fb posts, as a result of inoffensive representations flow into freely. The massive-scale manufacturing of AI-generated far-right content material, saturating the web setting, is more and more turning into mainstream. Whereas using far-right extremist symbolism is frowned upon or is even unlawful in sure nations, newly generated content material based mostly on ideological backgrounds is perceived as delicate when not depicting violence immediately. Nationalist teams and political events disguise racist, homophobic, patriarchal meanings in softened visible narratives to seem, not less than on the floor, normative and surpass digital platform restrictions.
When the integrity of political actuality is damaged down piece by piece, beginning in small communities via seemingly harmless channels, the way in which is paved for comparable ways to affect nationwide elections with out scrutiny. Political actors leverage using AI past content material technology, gathering information to create detailed psychological profiles of the citizens for extremely customized messaging. Covert operations producing giant quantities of AI-generated content material additionally enable politically motivated creators to see which matters and approaches resonate greatest. Occasional concern mongering is taken into account honest play, particularly previous to voting, simply to remind everybody of who supposedly public enemy primary is. However satire and different makes use of of humour are additionally proving to be significantly efficient.
How amusing are deepfakes?
Caricatures are outstanding in generative AI, which allows the manufacturing of infinite variations of individuals arrange in imaginary situations. The manufacturing of deepfakes based mostly on, for instance, photos of actual politicians’ faces has sparked debate in regards to the moral boundaries of such visible manipulation.
Throughout the 2023 election moratorium – the legally prescribed interval for campaigning in Slovakia – a deepfake scandal hit the web: a recording faking a telephone name between Michal Šimečka, chief of the Progressive Slovakia celebration, and journalist Monika Tódova initially unfold via Telegram. What was stunning only a few years in the past, nevertheless, has since grow to be an on a regular basis a part of the social media ecosystem. Extra lately, a deepfake video of Slovakia’s present coalition politicians and an oligarch supposedly in jail made by an Instagram consumer has additionally gone viral – partly resulting from it being shared by opposition politicians.
Whereas predisposed to deriding their opponents, following these examples and new EU laws, which got here into impact firstly of August 2026, Slovakian politicians are tackling deepfake movies. A full ban plus punishment starting from a positive to a possible jail sentence for as much as two years for offenders are within the offing. Consultants have in the meantime raised considerations about assessing accountability: not solely questioning whether or not those that immediate AI or the fashions themselves are at most fault but in addition recognizing the issue of discovering the precise creator of dangerous content material. In some circumstances, it’s clear who posted a picture, like on the Brat za Brata Fb web page. However many accounts are nameless and AI-generated content material is usually unfold via bot networks, eradicating culpability a stage additional. In Europe, content material distributed on giant on-line platforms is the duty of the Digitial Companies Act, however its operation is as a rule reactive, specializing in social media platform accountability fairly than on regulating the AI instruments used to generate deceptive content material.
High row, from left to proper: Matúš Šutaj-Eštok (minister of inside), Robert Fico (prime minister).
Backside row, from left to proper: Peter Pellegrini (president), Pellegrini and Juraj Blanár (minister of international affairs), Zuzana Čaputová (former president), Čaputová and Ľuboš Blaha (Slovak MEP), Peter Pellegrini. Picture by way of madeinslovakai Instagram account.
Proposed laws for safeguarding digital id could intention to guard folks from having their face misused for deepfake content material, however it addresses solely a part of the difficulty. It doesn’t start to deal with the overflow of AI-generated imagery throughout the political spectrum within the first occasion. Neither does it confront ‘believable deniability’ invoked to evade duty for offensive behaviour: ‘If somebody accuses you of, say, racism or homophobia, you may merely declare it was a joke,’ says Koblížek. And it doesn’t ask whether or not generative AI could be as efficient if it weren’t for current prejudices.
Reviving reactionary modernism
‘When you have a look at how the German weekly tabloid … Der Stürmer depicted Jews within the Nineteen Thirties, it is vitally much like how the AfD depicts migrants in the present day utilizing AI-generated photos,’ feedback researcher of political iconography Jacopo Galimberti. Nazi photos of idealized Aryan households may very well be equally in comparison with these produced by Brat za Brata. Stereotyping is a typical nationalist device that’s particularly efficient in visible type, exaggerating the variations between the ‘best’ and ‘the opposite’.
A lot Nazi propaganda relied on figurative representations of topics resembling the normal household, using recognizable depictions devoid of visible complexity and but bursting with civic delight. Machines had been equally glorified. The Nazi’s enthusiasm for technological progress was targeted on stopping the supposed decline of Western tradition and calling for a type of religious awakening. Submit-war, Jeffrey Herf termed this imaginative and prescient ‘reactionary modernism’, a motion able to paradoxically apply technological improvement in a bid to return to conventional nationalist myths and concepts.
Cowl of the ‘Neues Volk’ month-to-month journal calendar from 1938. Picture by way of Wikimedia Commons.
Latest developments in digital know-how, together with AI instruments and their makes use of, present parallels to twentieth-century reactionary modernism. ‘Technofascism’ describes the ideology behind technological enlargement over the democratic infrastructure of the state, empowering Huge Tech firms into authorities decision-making roles, beneath the course of the likes of Peter Thiel and Elon Musk. The aesthetics of Tesla’s cybertrucks and ‘trad wives’ come worryingly near the Nazi regime’s obsessions with army know-how and racially pure households.
What has modified, nevertheless, is the participatory facet of generative AI, which makes a content material producer out of each client. Nazi propaganda was orchestrated for the lots fairly than by the lots. Joseph Goebbels noticed movie as the principle medium for speaking Nationwide-Socialist ideology, using its leisure issue to additional wartime propaganda. Films from the interval resembling common comedies, whereas seemingly telling apolitical tales, painting racist insurance policies. Movie’s affect remained so key to the regime that even a number of months earlier than Germany’s capitulation, folks had been nonetheless being ‘drafted’ as extras for the Nationwide Socialist’s last-minute productions.
At the moment’s generative AI goes a step additional than participating civilians as extras on the massive display screen. It presents customers full participation and customization throughout content material creation, giving them management over their visualizations on the small display screen. Being free to make, AI-generated photos feed into grassroots types of propaganda. Now, provocative content material can go viral, assisted by automated bot networks boosting posts on customers’ feeds, disguised as clickbait. Members of Brat za Brata from Revúca, a city of 10,000 folks in central Slovakia, don’t want a producer or an area viewers. In a few clicks, they’ll visualize all of their concepts on id and energy constructions, and talk them to a community of customers who readily eat nationalist messages.
The enshittifying actuality of platform realism
Comparatively new but already pervasive, AI-generated on-line content material, produced in a vibrant and vibrant pallet with a recognizable visible attribute, has already earned itself a scholarly time period: ‘platform realism’. Coined by Jacob Birken, and additional developed by Roland Meyer, the idea identifies generative AI’s traits of legibility, plausibility and structurally nostalgic relationships to previous media. Meyer highlights how the very origin of generative AI merchandise separate them from actuality as we all know it, beginning with how LLMs course of the prompts they obtain. As Meyer explains: ‘regardless of the recognition of the time period “immediate engineering”, their perform shouldn’t be that of an actual program to be executed however that of fuzzy parameters narrowing down potential outcomes and steering the mannequin towards a selected area inside a seemingly infinite statistical area of potentialities.’ In the long run, the consumer has to strive a number of instances to get a picture that they like, which is inherently a wasteful course of each when it comes to assets and littering the web area with visible waste, nostalgia and clickbait.
In making an attempt to analyse this new actuality, researchers throughout many fields, from philosophy and sociology to research of processing and info methods, construct on Jean Baudrillard[i]’s idea of simulacra and hyperreality. Baudillard’s idea, which describes how representations can really feel extra actual than actuality, takes Disneyland as an illustration of how hyperreality replaces the truth past the theme park. He argues that we’re even previous that time: the ‘actual world’ is now stuffed with simulacra, making it nearly not possible to seek out ‘actual’ representations anymore.
The brand new-level simulacra mannequin is an autonomously generated actuality, approximated from our information archives. The creation of artwork, or on this case, the creation of propaganda, beforehand requiring human creativeness, has mutated right into a diminished sequence of prompts for AI fashions. And the representations that had been purported to persuade us of the true world that surrounds us have been changed by absolutely synthetic representations.
In keeping with thinker Marshall McLuhan’s media idea of know-how as ‘an extension of consciousness’, AI seems to offer members of Brat za Brata with their dream actuality. The group’s AI-generated imagery could properly even regulate their sense of self – assuming that these aesthetic decisions form their perceptions of actuality. As philosophers Philippe Lacoue-Labarthe and Jean-Luc Nancy focus on of their 1990 essay ‘The Nazi Delusion’, the ‘legendary energy is the ability of the dream, of the projection of a picture with which one identifies.’ Whereas the instruments used to painting and talk id have modified dramatically, the previous fantasy of Slavic unity carries on, rising in infinite exaggeration because of teams resembling Brat za Brata. However the creation of images with seemingly infinite potentialities comes with distinct caveats.
‘Machine studying works by correlation, by analogy, in a far more like alchemy than cause-and-effect-based science,’ says digital media artist Hito Steyerl. ‘There’s a proximity between magical considering and components of machine studying. However this magical vibe is blatantly pushed by giant companies simply to flog their merchandise; they’re not magical in any respect, however fairly enshittifying.’
Strategic air pollution of LLMs
LLMs be taught from the largely Western-dominated digital archives we offer them with. As soon as patterns have been acknowledged, they are often reproduced on a mass scale, in infinite variations. The prevalent info is subsequently biased.
‘Whenever you cease to consider the difficulty of duty, there may be usually a bent to shift the main focus onto the AI instruments themselves. We discuss what AI has introduced us, whether or not AI denies the Holocaust or not. Nonetheless, the duty really lies with the creators of the particular content material or the suppliers of the AI mannequin – and these obligations can fluctuate,’ says Koblížek.
Expertise strikes sooner than regulation. The speed and complexity of LLM improvement is such that it has already grow to be not possible to maintain up with whether or not the fashions are being skilled to unfold a sure rhetoric on function or if the folks in cost have failed to forestall these biases. The web area is being shared between common customers and inauthentic web sites mass-producing content material whose target market is different bots for future content material manufacturing. And, as Koblížek feedback, far-right teams are eager for his or her materials to grow to be a part of the LLM panorama: ‘Extremist teams actively work to make sure this content material is on the market on-line – even when it doesn’t garner many views – in order that it’s there as a possible supply for AI coaching.’
Given the way in which generative AI features, it additionally tends to strengthen the sexist and racist imagery, already current inside the digital archives feeding LLMs, which far-right actors are looking for. ‘It thus permits the fitting to conjure up a nostalgic, timeless previous that by no means existed exterior of the pictures used to coach the software program,’ says Meyer. The reciprocal mechanism of provide and demand is in place. Add a format that’s simple to share on-line, and a propaganda machine is born.
In some circumstances, we see politicians with different political leanings additionally leap on social media traits, producing their very own content material that mocks their opponents. Nonetheless, ‘liberals, progressives and leftists, in contrast, have largely been reluctant to publicly use AI – not solely due to the fashions’ well-known biases but in addition due to the huge exploitation of human labour and creativity, in addition to of pure assets concerned in coaching AI fashions. For the fitting, then, embracing recognizable AI or “slop” aesthetics can also be a approach of triggering and trolling their opponents,’ provides Meyer.
Quickly we could have discussions not solely about producing new photos but in addition about defending the integrity of the previous, which, in an Orwellian approach, can simply be altered. The US authorities has already began by erasing photos of ladies, LGBTQIA+ folks and folks of color from its army web sites. Meyer warns of ‘digital forgetting’ – in different phrases, we shouldn’t be so certain of the ‘complete availability of the previous’. ‘This not solely breeds a pervasive mistrust in visible proof of any variety but in addition produces deceptive expectations about what historic paperwork ought to appear like,’ says Meyer. Once we erase the originals from our archives, they might properly solely stay as shadows within the coaching materials of LLMs.
Museums and public establishments typically embrace generative AI themselves, to fill in historic gaps and produce the previous into life. In keeping with Meyer, ‘such speculative makes use of are sometimes deceptive. They have a tendency to breed current clichés in regards to the previous and the way it needs to be represented, fairly than opening up a brand new view of historical past.’
Establishments are confronted with conserving concepts of reality alive, growing public consciousness and serving to outline codes of follow on AI use. Regulating on the velocity of growing know-how can be a relentless problem.
In the meantime, AI-generated content material from teams like Brat za Brata that may maintain being produced advert infinitum could properly attain a degree of saturation, finally solely persuading the persuaded or just tiring its viewers out – as with different types of propaganda earlier than it.
This text is printed with the help of the ZEIT-Stiftung Ebelin und Gerd Bucerius.
