The Westerplatte memorial in Gdańsk on Poland’s Baltic coast marks the site where WWII began with a naval assault in September 1939. The day I visited in September 2025 happened to be the same day that Russian drones were shot down over Poland, marking Moscow’s first overt incursion into NATO territory since its full-scale invasion of Ukraine. Alongside all the usual tourist amenities – toilets, snacks and souvenirs – I found every kiosk lining the memorial ground stocked up on plush toys with some unexpected faces: Tralalero Tralala, Ballerina Cappucina and, of course, the unmistakable oblong shape and leering eyes of Tung Tung Tung Sahur.
This was Italian brainrot, a pantheon of AI generated characters who appeared on social media in early 2025 before seemingly taking over the world for the next few months. A shark wearing three blue Nike trainers, a dancer with a coffee cup for a head, and a humanoid stick carrying a baseball bat emerged as a few of the favourites. Each character had its own pseudo-Italian name, musically pronounced in each video by the same AI text-to-speech voice tool, ‘Adam’.
Seeing these terminally online memes turn up in polyester, here, of all places, gave me an unsettled feeling. Italian brainrot for sale at a Polish memorial to victims of the German Third Reich. On the day that the post-war military order frayed at its edges. It felt like irreconcilable worlds coming together, digital native nonsense spilling into real life, flooding a site of meaningful remembrance, and renewed significance. It was surreal, a waking dream. This is a feeling that novelist Thomas Pynchon named ‘virtuality creep’ in 2013’s Bleeding Edge: when the digital realm overflows into the ‘perilous gulf between screen and face’.
Monument to the Defenders of the Coast, Westerplatte, Gdańs
Italian brainrot for sale, Westerplatte, Gdańsk
Most internet memes are images wrenched from their original context (TV shows, cultural and political events etc.) then endlessly remixed and repurposed, their meaning flattened in the infinity of the web. With Italian brainrot, the reverse was true: these avatars emerged from the digital ether with absurdity pre-loaded. They never meant anything. Early clips featured Italian ‘lyrics’ for each character which added little context. Some were cryptic origin stories while others were just expletive filled and crassly offensive rants. The clips featured a patchwork of recognisable features which invited interpretation – animals, objects, environments, a vague notion of Italianness itself –but together amounted to nonsense.
Upon going viral, the characters were quickly plucked from their social media habitat and put to work in our world: in cartoons, songs, video games, classrooms and the cheap tat for sale at every stall in Westerplatte. After these bizarre chimeras had captured people’s attention, wily content creators and the machinery of consumerism each seized the opportunity of new IP with untapped potential, in pursuit of clout or commerce. The characters were simple enough to entertain children, weird enough to mystify parents and eye-catching enough to entice the algorithm.
“And the Oscar goes to… Brr Brr Patapim”
Brainrot content is so exaggeratedly fake, the funniest thing to do quickly became obvious: act like it’s real. The initial emergence of Italian brainrot on social media was followed by a wave of painstakingly detailed, completely contradictory, mostly hilarious ‘lore’ videos. Straight-faced explainers recounted character backstories which sometimes sounded like twisty soap opera storylines, like Ballerina Cappucina’s husband Cappuccino Assassino being seduced into an affair by love rival Espressona Signora. Other characters were shrouded in esoteric myth, like Lirili Larila, the most Daliesque of the lot, who has an elephant head with a cactus body and one leather sandal. In most accounts, Lirili Larila is a kind of lonely god, wandering the desert and cursed with the ability to manipulate time itself.
Some lore expanded details from the original Italian lyrics, while others started from scratch. These AI avatars were not-quite-blank canvasses, an evocative visual language ready to be loaded with significance, from the sublime to the ridiculous. Mostly the ridiculous.
My feed was overtaken with videos pitting the characters in battles against one another, alongside deepfakes of King Charles naming his champion as ‘knight of the realm’ or AOC decrying the power of ‘big corporate sponsored fighters’. This content increasingly bled into the real world. I watched Cillian Murphy present an Academy Award to a brainrot character. Creators published live-action videos in the style of a nature documentary or a found footage horror, with titles like “IF YOU SEE LIRILI LARILA WHILE DRIVING, DO NOT APPROACH…😱’ and ‘DRONE CATCHES TUNG TUNG TUNG SAHUR ACROSS MY CITY!! (SCARY)’. The trend settled on treating Italian brainrot characters as elusive but real beings.
One of the most popular formats was street interviews, usually involving a young person showing older strangers pictures of brainrot characters and asking if they know who it is: “Ahhh yes, that’s Crocadilo Bombardiro” they reply, deadpan. Instead of plugging our own likeness into AI, to make lame Studio Ghibli imitations or action figures, these prankish videos dump AI images into the physical world. The humour is in the public recognition, brainrot characters unexpectedly recalled like an old friend from school.
“Everyone knows that!”
This same logic, of the virtual absurd intruding into mundane reality, gave life to the popular ‘John Pork is calling’ trend, where creators share videos of people receiving an unexpected phone call from an anthropomorphic human-pig influencer, his uncanny grinning face flashing up on screen next to the ‘accept’ or ‘decline’ buttons. It’s a goofy ahh Gen-Z reimagining of the ‘red pill, blue pill’ conundrum from The Matrix: will they reject the call and end the story, or pick up and see how deep the brainrot rabbit hole goes.
Death of the author – who owns AI folklore?
AI generated characters, like the Italian brainrot crew and the cast of the John Pork lore, make possible new kinds of storytelling. These figures seemed to appear out of nowhere, largely without claims of ownership and easily reproducible with AI tools, which invited creators to run with it. The John Pork lore evolved into a kind of mass-participation true crime investigation to unravel John’s supposed death. Countless videos on TikTok or Youtube purported to reveal the ‘complete lore’, the ‘real story’ or – my personal favourite – the ‘dark truth’ of John’s disappearance. Often cloaked in a conspiratorial film noir style, this is a search for meaning at its core: ‘the story you think you know, it’s just the tip of the iceberg’.
This anarchic din of narratives crucially existed on a level playing field; no-one had a monopoly on a definitive version of events. Brainrot lore oozed across social media as a rhizomatic sprawl, without any kind of central authority. Attempts to file for copyright and trademarks over Italian Brainrot characters have largely been frustrated by legal gray zones around questions like whether an AI prompt constitutes original authorship, and if AI generated works are entitled to copyright protection. In March 2026, the US Supreme Court delivered a blow to brainrot privatization by declining reconsider a district court ruling that human authorship is the ‘bedrock requirement of copyright protection’.
The current wild west of AI generated content goes against an age of aggressive corporate takeovers and nostalgia-obsessed cultural production, when studios covet intellectual property and write up convoluted loan agreements over who can depict which superheroes and for how long.
The flourishing of brainrot lore echoes a much older kind of entertainment: cultural folklore forged from tall tales, yarns, gossip and rumours, reshaped by anyone with the imagination to tell it. User-generated brainrot lore handles AI with mischievous glee, chaotic creativity and a hunger for shared meaning-making. It’s a messy democratisation of storytelling. These jumbled narratives highlight the intertextuality of meme culture; by refuting or retconning details of the lore, each intervention widens the story further while reinforcing the idea attaining a singular, concrete truth.
‘It wasn’t a conscious pushback against platforms or their algorithms’ argues cultural theorist Daniele Zinni in a recent Brainrot retrospective for the Institute of Network Cultures, “it was more like a revival of the internet’s anarchic prankster spirit, which periodically reemerges’: the fun came from ‘collectively giving attention, for as long as possible, to something so trivial.” The underlying principle is total commitment to the bit, an unwinking verisimilitude no matter how silly things get. It’s no laughing matter when Bob Bacon and Marvin Beak are so close to avenging John Pork’s death at the hands of Tim Cheese. Or perhaps you side with contingent of sleuths who insist that Pork is not dead at all, and that Cheese is an innocent AI-generated anthropomorphic mouse. Deadly serious stuff.
@money.universityy, Instagram
Unreality bites
The impulse to treat brainrot with total sincerity can be read as a playful outlet to a deeper feeling which is creeping into the rest of our lives, particularly for young people. This is a sense of unreality, a nagging sensation that everything is fake now. In so many ways, it’s getting harder to separate reality from fiction, or maybe it’s just pointless. We increasingly inhabit a kind of third-person POV, consciously curating our life as we do our feeds. ‘Authenticity’ is nothing more than a marketing strategy. The rise of unscripted ‘IRL streams’ renders public space a content creation studio. We talk incessantly of ‘performative’ behaviour, NPCs, ‘main character syndrome’, ‘doing it for the plot’. The real has never been so entangled with the virtual, experience never so inseparable from representation.
Marxist theorist Guy Debord sensed early on how an image-saturated culture would shape the way we experience the world and relate to each other. ‘When the real world is transformed into mere images’, he wrote in 1967, ‘mere images become real beings’. From social media and fitness trackers to smart doorbells and live-feed CCTV screens in supermarkets, we are trained with increasing intensity to grasp life through a dialectic of watching and being watched, always playing to an audience both real and imagined. As Britney Spears said: ‘there’s only two types of people in the world, the ones that entertain and the ones that observe’. In the endless refraction of surveillance capitalism, these types have been folded into one.
British speculative fiction author J.G. Ballard reflected in 1990 that a ‘huge volume of sensational and often toxic imagery inundates our minds, much of it fictional in content’, asking ‘how do we make sense of this ceaseless flow of advertising and publicity, news and entertainment, where presidential campaigns and moon voyages are presented in terms indistinguishable from the launch of a new candy bar or deodorant?’. Ballard was of course talking about television and film, but 50 years later the same question lingers over TikTok, Instagram and our short form video addiction. “What actually happens on the level of our unconscious minds when, within minutes on the same TV screen, a prime minister is assassinated, an actress makes love, an injured child is carried from a car crash?’. We’re still figuring it out. At this moment, our inner space is being rewired by a rapacious media matrix, while our outer world – the ‘real world’ – is being revealed as a collective delusion.
@yungstarbeam, Instagram
Nothing is true and everything is possible
From the pandemic’s global shutdown to the carnage in Gaza and Ukraine, we are living through a dizzying collapse of the old certainties – ideas of security, order, truth and justice – which anchored our sense of reality. The neoliberal exaltation of the market and the fanaticism of the populist right betrays a politics of zealotry. Our ethical sense is overwhelmed as we consume ethnic cleansing as online content and the perpetrators claim victimhood. The Epstein files disclose a cosy conspiracy of transnational elites united by depravity and impunity. Each passing day pulls back the curtain on the world of rules and rationality that we were taught to believe in. It’s a reverse Wizard of Oz where instead of the mundane masquerading as magic, we are gaslighted by absurdity cloaked as sober reality.
And yet nothing really changes. We know the world is burning and the system is broken, yet daily life continues as normal. We go to work and the gym and the shops. This is where the dissonance creeps in. As critical theorist and content creator Louisa Munch recently put it: ‘every day we are performing belief in a system no-one believes in’. This is where I find a subversive streak in brainrot lore: this content is also a ‘performance of belief’ but a conscious one. In its ridiculousness – performing belief in something patently unbelievable – it calls into question the other ways we perform, and suspend, our belief. In a time of mass-cynicism, credulity can be wielded as a scalpel… or a baseball bat.
Disneyland, provocatively claimed French philosopher Jean Baudrillard in his seminal work Simulations and Simulacra, is ‘presented as imaginary in order to make us believe that the rest is real’. It is an exaggerated fantasy which serves to reinforce our belief in the rest of our everyday reality, which in fact now exists only within a procession of images and illusions: ‘the hyperreal order and the order of simulation’. As a performative act, Brainrot lore flips this on its head. Tung Tung Tung Sahur presented as real reminds us that the rest is imaginary.
‘When the world becomes unintelligible, humour grows teeth’ says researcher and UX designer Moreno Nourizadeh, ‘the surreal is always a revelation of the real’. Look back and we see that brainrot (and its discontents) is nothing new. Every generation, writes Nourizadeh, succumbs to an ‘epochal narcissism’; this is the conviction that ‘its particular madness is unprecedented, that its stupidity signals unique decline’. When Lewis Carroll’s nonsense literature poked fun at rigid Victorian hierarchies and the logic of language, literary magazine The Athenaeum wondered if Caroll had ‘merely been inspired to reduce to idiocy as many readers as possible’. The anti-rational Dada art movement grappled with the civilizational impact of WWI’s industrialized slaughter, and met with disdain and disgust.
Brainrot as a cultural form and the set of lore practices which emerged around it, is a product of the AI revolution, a barely processed pandemic and the collapse of the post-1945 world order. It is a contradiction: lore is about shared meaning-making and assembling pieces into recognisable narrative shapes, brainrot is about revelling in nonsense. It pokes fun at our current epistemic crisis, where we are losing our grip on reality itself. Chat, is this real? Is this Large Language Model my friend?
Hannah Hoch, Man and Machine, 1921
Lirili Larila, 2025
Italian brainrot presents the AI generated image as something excessive, nonsensical and deranged. This, at the same time as Silicon Valley companies are spending billions to naturalize AI and weave it imperceptibly into our daily routines. On every app now we are accosted by AI tools that no-one asked for and no-one can remove. Brainrot introduces a glitch in the AI matrix, a violent jolt in the frictionless world. Remember, the half-frog, half-tyre character Boneca Ambalabu is just as real as the photorealistic AI ‘slop’ flooding social media. This ranges from impossibly curvy OF girls to ‘footage’ of racialized ragebait or sympathy-fishing images of crying children and puppies rescued from floods. You’re not meant to think it’s real, just feel that it could be.
In a taxonomy of AI content, the intentional absurdity of brainrot is its defining feature. It invites interpretation then laughs at our efforts. Brainrot content may be AI generated, Daniele Zinni argues, yet it is ‘anything but statistically probable’ and ‘strange enough to surprise rather than trigger a predictable reaction’. It’s hard to feel that Tralalero Tralala is real, hence the satirical mileage in pretending that he is. Slop is more insidious. It wants something from us. Emotionally coercive slop images leverage our attention for further motives, trading in our prejudices, allegiances and desires. It is the tool of the trade, writes Gunseli Yalcinkaya, for ‘grifters trying to make a quick buck’ and ‘politicians wanting to overwhelm the system with AI cringe edits of themselves as Star Wars characters’. Slop is the medium of narcissistic wish fulfilment, so it makes sense the current White House cannot resist its lure.
Merriam-Webster’s 2025 word of the year, much has been said of brainrot, both as form and function. It is often used interchangeably with slop in public discourse. This mind-numbing material will supposedly destroy young people’s critical thinking skills, turning them into babbling zombies. What if this bleak prophecy was true, but not for us? A slop-heavy digital landscape of fake news, goonbait and synthetic influencers might contain a fatal trap for the AI revolution itself.
The shift to an AI-mediated internet, Kate Crawford argues in e-flux, is shutting out human content creators, who face ‘face dwindling readership, reduced engagement, and disappearing ad revenue’. Yet multiple studies have shown that AI systems ‘degenerate when they are fed on too much of their own outputs’, a phenomenon researchers call MAD (Model Autophagy Disease). In other words, Crawford writes, ‘AI will eat itself, then gradually collapse into nonsense and noise’. Enshittification tolls for thee. In this sense we can see Italian brainrot as an accelerationist provocation: its absurdity simply pushes the current AI regime to its logical conclusion. It’s the demented future waiting in the code. Chat GPT looks in the mirror, nervously, and sees Tung Tung Tung Sahur glaring back.
Triple T
Seeing was believing
I would argue that it’s not the absurd material produced by AI that we should find repulsive, but anything that we might take, at first glance, as real. In just a few years since we laughed at the nightmarish sketches produced by early image generator DALL-E, we now have tools on our phones, like Google’s Nano Banana Pro, which can conjure entirely convincing synthetic people with the right amount of fingers. This effectively means that photography as means of gathering evidence, documenting real events is finished. Instead we have an exhausting cynicism where every image should now be interrogated with scepticism. Footage which once might have stirred public outrage – of corruption or war crimes – loses its moral authority, dismissed with the wave of a politician’s hand.
AI Will Smith eating spaghetti, @aiemerges, Instagram
It’s not just that AI images undermine the veracity of images, but the troubling new ways they can simulate images. Genealogy site MyHeritage offers (for a monthly fee) to ‘bring dead ancestors back to life’ by using deepfake technology to animate their faces in photographs. The company introduced the ability for these images to speak, just a year after assuring they wouldn’t. Grief Tech is an unsurprisingly fast growing market, because who wouldn’t want to reach out to a lost loved one? Researchers at the University of Cambridge have called for greater guardrails, warning how these deadbots may create unhealthy emotional attachments, stifling the mourning process and making the bereaved prone to manipulation. Imagine hearing your dead grandparent tell you that your monthly subscription price is going up. It’s multiple Black Mirror stories at once.
In a sense, AI image generation is the terminal endpoint of the centuries-long Enlightenment project of illumination. This was the idea that everything in darkness should be dragged into the light, the invisible made visible, scientific reason as the means and the end. The principle of empirical observation and transparency underscores everything from democracy and secularism to mass-media and surveillance. Now AI image generation gifts us mortals the ‘God’s eye view’, all-seeing and therefore all-knowing. We can make whole worlds in 7 seconds. ‘There are eyes everywhere. No dark spots left’ once remarked cultural theorist Paul Virilio: ‘what will we dream of when everything is visible? We’ll dream of being blind’.
The ability to speculate, anticipate, to fill in the blanks, defines our human relation to futurity, to joy and to beauty. ‘Heard melodies are sweet, but those unheard are sweeter’ wrote poet John Keats. “Where do you see yourself in five years” we ask ourselves. Our mind’s eye is a hazy, flickering image we must actively grasp for but can never hold. Until now. Our desires wax and wane like the moon, AI crystallizes them into solidity. Technology is marching on one of the last dark enclaves, our imagination, to steamroll into a flat commodity. Brainrot is dreaming’s revenge.
—
George Harry James is a London-based writer and cultural critic whose work explores digital aesthetics, online subcultures and what it means to live in the world today. He has a Masters in English Literature from University of Sheffield. You can reach him at georgeharryjames1@gmail.com
—
References
– The Athenaeum Journal of Literature, Science, the Fine Arts, Music and the Drama, January to June, 1876.
– J.G Ballard, The Atrocity Exhibition, London: Jonathan Cape, 1970.
– Jean Baudrillard, Simulations & Simulacra trans. Sheila Faria Glase, Ann Arbour: University of Michigan Press, 1994.
– Francesco Barchiesi & Geert Lovink, The Story of Italian Brainrot – Collective Musings on a Meme Wave, Institute of Network Cultures, 30 September 2025, The Story of Italian Brainrot – Collective Musings on a Meme Wave.
– Kate Crawford, ‘Eating the Future: The Metabolic Logic of AI Slop’, e-flux, September 2025. Intensification – Kate Crawford – Eating the Future: The Metabolic Logic of AI Slop.
– Guy Debord, Society of the Spectacle trans. Ken Knabb, Berkeley: Bureau of Public Secrets, 2014.
– Tomasz Hollanek, Katarzyna Nowaczyk-Basińska, ‘Griefbots, Deadbots, Postmortem Avatars: on Responsible Applications of Generative AI in the Digital Afterlife Industry’, Philosophy & Technology 37, May 2024.
– John Keats, ‘Ode on a Grecian Urn’ The Poetry Foundation,
www.poetryfoundation.org/poems/44477/ode-on-a-grecian-urn.
– Jacob W. S. Knight, ‘The Final Word? Supreme Court Refuses to Hear Case on AI Authorship and Inventorship’, Holland & Knight, 3 March 2026, The Final Word? Supreme Court Refuses to Hear Case on AI Authorship and Inventorship | Insights | Holland & Knight.
– Louisa Munch, @louisamunchtheory, Neoliberalism is over pass it on, Instagram, 2 March 2026.
– Moreno Nourizadeh, The Phonemic Flesh: A Phenomenological Analysis of Italian Brainrot, Zenodo, 26 February 2026.
– My Heritage – Animate your family photos, My Heritage – Animate your family photos, MyHeritage Deep Nostalgia™, deep learning technology to animate the faces in still family photos – MyHeritage.
– Thomas Pynchon, Bleeding Edge, London: Jonathan Cape, 2013.
– Britney Spears, ‘Circus’, Circus, Sony Music Entertainment, 2007.
– Louise Wilson, ‘View of Cyberwar, God And Television: Interview with Paul Virilio’, CTheory, October 1992.
– Gunseli Yalcinkaya, ‘Digital Dada or Futurist slop? An investigation into brainrot as art’, Plaster Magazine, 11 June 2025, Digital Dada or Futurist slop? An investigation into brainrot as art – Plaster Magazine.