V-AI-mpire: Media monsters
This article is also available in French.
Anyone expecting Radu Jude’s latest film to be a new “biopic” about the bloodsucking creature will be disappointed. In fact, its nested narrative follows an uninspired filmmaker (Adonis Tanța, playing a histrionic stand-in for Jude himself) who turns to AI to create a new incarnation of Dracula. His original film features two actors dressed as vampires fleeing a grotesque tourist-trap show somewhere in present-day Transylvania. But the machine soon runs amok, generating an array of wildly different proposals – all of which the director dismisses as unfinished, absurd, or indecent.
This back-and-forth between creator and machine unleashes a vast parade of possible vampires, drawn from all sorts of spatial and temporal dimensions: classical monsters and modern avatars, Western or local archetypes, from Nosferatu to Vlad the Impaler, with a few unpredictable Romanian references thrown in. Their degrees of embodiment vary just as much: supernatural entities mingle with costumed serial killers, old men obsessed with rejuvenation, disguised actors, or metaphorical images of economic and sexual ravage – like those phallic corncobs impaling everything in sight. Even the vampires themselves end up vampirised: transformed into tourist attractions, exploited sex workers, or pitiful devils with wounded mouths at the mercy of greedy dentists.
The myth’s unstable, shape-shifting identity is further amplified by its cross-media nature: from literary beast to graphic monster, from cinematic spectre to social-media chimaera. As the narrator warns us, Frankenstein might have been a more fitting title for such a composite monster of a film. In short, in Jude’s Dracula, the vampire’s identity matters less than the process of vampirism itself. The vampire is embodied not so much by any character, but by the film itself. In other words, the vampire is the media. But what does a vampire medium mean?
A Greedy Medium
A vampire medium is, first and foremost, a greedy one. Over nearly three hours, Jude’s film devours a hemorrhagic flow of references, images, and media. The monster libido – represented quite literally through a sequence of erotic images – fuels the project’s iconophagic frenzy. After all, the vampire is an obsessive creature, driven by a devouring hunger that is both ritualised and feral. And since this film is a vampire, it activates itself by compulsively feeding on the flow of its own media history.
Naturally, that means the history of horror cinema: from Nosferatu (1922) and German Expressionism to Dreyer’s Vampyr (1932), through B movies, Hammer’s blood-red productions, Doru Năstase’s Vlad Țepeș (1979), zombie flicks, and contemporary snuff aesthetics. But the film’s digestion goes even further, swallowing up everything from slapstick and Fluxus avant-garde to Soviet realism, which echoes in its rural-realist and robotic-futurist sequences.
The AI program used by the filmmaker character offers a clue to the nature of this machine: dubbed DR AI Judex 0.0, the interface nods to Louis Feuillade, creator of the silent serials Fantômas (1913), Judex (1916), and Les Vampires (1915). These popular feuilletons – brimming with erotic charge during the Great War – helped define cinema as an addictive substance, toying with viewers’ desire to return to the screen day and night as new episodes are shown. An inexhaustible spectacle – precisely the principle behind Jude’s expansive montage, which piles up chapters, digressions, and excerpts.
But even cinema’s potentially infinite reel isn’t enough to sustain the beast. Or rather, cinema – that old machine built to absorb cultural referents – is now itself parasitised by an even greedier organism: AI, a technology that consumes millions of archives to feed its own calculations. The analogy between vampirism and technological enhancement is established early on, when the film quotes Ludwig Eduard Boltzmann, the pioneer of statistical physics. According to a probabilistic view of ghosts, the more the dead multiply, the higher the chance that one of them will return to life. Jude’s update of that idea is nauseatingly contemporary. The inexhaustible has become exhausting. The scopic drive associated with 21st-century porn culture, which the film lays bare, seeks – pardon the expression – to plug every hole, to fill every gap in its proliferating images. It strives to saturate and colonise the imagination itself.
(Un)Dead Media
A vampire medium, by definition, accumulates and digests time. Yet its supposed eternity constantly collides with anachronism. The gap between the ancient monster and the eras it inhabits runs throughout modern cinema: from the outdated romanticism of Coppola’s Dracula Dark Shadows (2012), where an ethereal vampire is confronted with consumerism and mistakes the glowing “M” of McDonald’s for Mephisto’s signature.
Paradoxically, today’s vampires are more like playthings than masters of time – like Vlad the Impaler in Jude’s film, who is chased out of his own home by a tourist guide. That’s because various media now serve as their bodies: from PowerPoint slides to commemorative gravestones.
These vampiric media linger among the ruins, i.e. the fading traces and vanishing archives. Jude’s film unearths these buried or obsolete genetic sources, the so-called dead or lost media: the waste and dumpsters of history, scavenged by marginal vampires. One striking example is “Dracula Park,” a 2001 project launched by the Romanian government that was never built, resurrected here as an abandoned webpage – a perfect non-place, both in the real and virtual world.
Alongside erasure, another form of ruin is degradation. Affected by Walter Benjamin’s famous “loss of aura” in the age of technical reproducibility, the film’s scenes follow a parodic trajectory of decay. Its low resolution – shot on an iPhone 15 – is integral to the operation. Among the many directions Jude has opened, we might note his subversion of funerary traditions. The Fayyum portrait in Do Not Expect Too Much from the End of the World (2023), detached from its site (the tomb and the body), was reduced to a meaningless flat surface on the stark white wall of a mortuary chapel. In Sleep #2 (2024), Andy Warhol’s grave – an artsy time capsule captured by online surveillance cameras – became a screen for necrological voyeurism. Here, the cemetery becomes the ultimate site of dark tourism, evoking the snuff-movie subgenre.
And yet, vampires – evolutionary creatures by nature – form complex palimpsests. Jude’s montage mirrors this accumulation of temporalities. Structured into chapters whose numbering skips and stumbles, the film follows a non-linear logic that derails from one segment to the next. The director-narrator’s initial thread – tracking two actors (Oana Maria Zaharia and Gabriel Spahiu) hunted by tourists armed with stakes – unfolds chronologically only to be ripped apart by the multitude of haunting pasts, edited alternately.
The narrator’s voice, meanwhile, both links and fractures the fragments, cutting stories short or splintering them with sardonic commentary. The “synthetic” images allow for maximum visualisation of this heterogeneous temporality: time itself seems to coagulate within them. Through their collage effect, they heighten the rift between figure and background – as in a poorly rendered 3D vampire pasted onto a faded Byzantine fresco.
Equally so, the series of still or stuttering images that disrupt the film’s flow – echoing the split screen between Murnau’s Nosferatu and its vulgar, ad-inflected appropriations – reveal the machine’s fundamental incomprehension of the past: its inability to compute or stitch it together. They also point to Jude’s Benjaminian choice – not to work with beautiful relics, but with the ugliness of the present.
Uncanny AI
Media theory has long emphasised the “uncanny” dimension that accompanies each new technology of reality reproduction. Jude’s film rekindles our fascination with illusionist machines – whether it’s the fairground spectacle or artificial intelligence itself. In many ways, generative AI revives the old art of prestidigitation. “Text-to-image” creation has something of a spell about it: the performative word seems to summon creatures on demand, instantly.
Outsourcing creative labour to the machine might undermine the virtues of personal effort and originality – and Jude fully embraces this laziness – but it also opens the door to a surprising, “non-human” sensibility. The fact that the machine’s computational process remains partly opaque even to the creators only adds to its allure. AI’s literal yet aberrant way of visually translating literary sources often produces astonishing effects. When adapting Nicolae Velea’s short story Just So, AI generates impossible images depicted in the text. It materialises what the story insists cannot exist – “there couldn’t possibly be rivers of milk one might drown in.”
The “life of objects,” even to the point of animism, so central to 1920s cinema, also finds new expression in these technologies. Techniques such as isolating the object from its background, prosthetic image regimes, and bizarre compositional effects give rise to strange scales and hybridities (including gigantic sex organs) and surreal juxtapositions. Human morphology dissolves into this monstrous plasticity: here, several artefacts merge into a chimaera; there, algorithmic montage pairs elements according to inscrutable logics. Something of the surrealist fetish – the random coupling of objects – is reawakened.
Jude seems to reconnect here with Gherasim Luca’s The Passive Vampire, which championed “objectively offered objects” (O.O.O.), products of “objective chance” that were disjointed and recombined, sometimes violently, to recover their unconscious and erotic charge. The latent life of objects is translated by synthetic imagery, particularly in its unfinished states – sketches, drafts, fragments that reveal a buried world of matter not yet expressed.
Even so, Jude’s film doesn’t renounce the more immediate emotion of recording reality. The return to realism amid this avalanche of artifice becomes, paradoxically, yet another layer of strangeness. The segment Just So, shot in the socialist-realist style of the 1960s, produces a comic yet unsettling effect through the weight of a camera mounted on a moving truck. The jolting motion of the vehicle gives the heroine’s chest (Alexandra Harapu) an erotic bounce, soon replaced by the twitch of a lifeless body. Similarly, the weary flesh of Gabriel Spahiu’s impotent vampire offers an affecting counterpoint to the phalocratic artificiality of the “augmented” ones.
Finally, the epilogue makes one last attempt to evoke an emotional version of the vampire legend – an emotion to which AI proves impervious. The mise-en-scène is disarmingly simple: the chromatic and spatial separation of a father, kept at a distance from the courtyard where his daughter plays, isolated by a fence and a fluorescent garbage collector’s uniform.
Far from limiting itself to an apocalyptic or demonising portrayal of new technologies, Jude’s Dracula instead maps a new balance of power. The constant push and pull between the director and AI exposes the tool’s aesthetic and ideological limits – its built-in censorship, its cultural stereotypes, its geographic biases. The director’s capricious dissatisfaction dominates the machine as much as he is dominated by it. And, as Jude suggests, it’s not at all certain that cinema won’t end up vampirising AI in return.
Élodie Tamayo
ÉlodieTamayo is a lecturer in cinema at École Normale Supérieure Lyon and film critic at Cahiers du cinema. She is researching history and mass-media theory and is obsessed of partial or unfinished works.
Élodie Tamayo
[:ro]Élodie Tamayo este conferențiar universitar în domeniul cinemaului la École Normale Supérieure din Lyon și critic la Cahiers du cinéma. Cercetează istoria și teoria mass-media şi e obsedată de operele parțial sau deloc finalizate.[:en]ÉlodieTamayo is a lecturer in cinema at École Normale Supérieure Lyon and film critic at Cahiers du cinema. She is researching history and mass-media theory and is obsessed of partial or unfinished works.[:]