Where do all the unwanted AI images go once they’ve been discarded? That’s the question I ask myself as I tap away at Midjourney’s text-to-image generator. We don’t know where these images come from or where they’re going—they’re just spat out, always appearing yet never quite there. These neglected entities—distorted, disembodied, deracinated—haunt the image graveyard, doomed to an eternity spent inhabiting hidden corners of the net. Enter the spookhouse at your own risk, I think, and press Enter.
The Extremely Online have long sensed the internet's haunted nature, but the recent wave of text-to-image AI has revealed all kinds of monsters lurking in the shadows. Take Loab: a macabre image of a woman with bloody cheeks and hollowed-out eyes that haunts every image she touches. Upon discovery last year, Twitter users quickly labelled her “the first cryptid of the latent space,” a term used to describe legendary creatures like Bigfoot or Mothman whose existence is disputed by science. More than simply being a made-up creature, however, the cryptid exists through collective belief: a patchwork of stories passed between users through traces, rumours, sightings and pixelated photos, evidence that is either deliberately faked or faked evidence that believers sincerely believe to be real. Though their presence is felt across the net—through hyper-dedicated subreddits and TikTok explainers—the ultimate reality of cryptids is undetectable. Loab herself was first “discovered” seemingly at random after a user named @supercomposite found that entering various text-to-image prompts and combinations repeatedly generated the same image, which has since spread across the net at warp speed.
Another AI cryptid is Crungus, a monstrous and hairy humanoid that bears an uncanny resemblance to the horned figure in Alpine folklore called Krampus. Though much less horrifying is Erosion Bird, aka the majestic creature from 2027—his Luh Calm Fit has inspired countless memes, video edits, and spaced-out Soundcloud clips. These creatures have the same contagion as human-authored creepypastas such as Slenderman and Momo. Perhaps unsurprising given how vast chunks of data scraped from the strange corpus of online culture are one of the most popular sources of training material for generative AI tools. At once sinister and playful—a haunting amalgam of our collective fears or simply a funny meme—they appear dreamt up from the recesses of the deep-net hivemind, an amalgamation of an internet’s worth of images selected, superimposed, and processed past recognition. An untraceable entity—digital cryptids have no origin story—they appear seemingly out of nowhere, growing out of the ever-evolving feedback loop of stories that surround them. These extend far beyond our computer screens into wider pop culture, where they’re chewed and digested and sold back to us in the form of video games or the latest A24 horror.
Whether real or imagined, these ghost stories and urban legends persist through time, across online forums and social media, where users post fried night cam edits alongside #cryptidcore clips of dark forests, small towns, VHS tapes—an aesthetic loosely inspired by paranormal cult shows like The X-Files and Twin Peaks. They feel as real as anything else online does; stories of cryptids are passed from user to user and transmitted into the Cloud, where they become part of our collective internet folklore, along with other similarly untraceable online-coded things such as girl dinners, subway girl, psy-ops, Pinkydoll, Julia Fox. Cryptids are avatars of the creeping sense of unreality that pervades our online experience—things of unknown origin that require us to consider what’s “out there.”
In his 1933 essay Some Notes on a Nonentity (1933), H.P. Lovecraft pinpoints the relationship between reality and the weird: “The punch of a truly weird tale is simply some violation or transcending of fixed cosmic law—an imaginative escape from palling reality.” Nightmarish entities lurking somewhere beyond the realms of our perception, cryptids are Lovecraftian. They stalk the stale yellow corridors of the Backrooms between the real and imagined, leaving behind a digital wormhole of memes, short-form explainers, and conspiracy theories. Imagining a cryptid asks us to look beyond rational thinking—the version of reality constructed under science and technology. This means encountering the unknown in a way that an obsessive need to understand—typical of our contemporary desire to shovel everything into categories—might block.
Anything that exists beyond the borders of our reality is weird—aliens are weird, algorithms are weird, monsters are weird, 5G is weird, Benjamin Bratton dressed in a lizard mask is weird. The cryptid similarly exists in this non-space—it’s undefinable and therefore challenges the categories and identities we use to comprehend it. In the Middle Ages, bestiaries containing detailed descriptions and illustrations of animals were taken from verbal accounts. These were often physiologically incorrect, making no distinction between existing and imaginary animals—think detailed descriptions of dragons, unicorns, and griffins dispersed between accounts of bears, deer, and lions. The meaning behind these texts is still debated by historians—were they meant as an actual compendium of creatures, a wildly illustrated morality tale, or pure entertainment? Current-day cryptids are similarly ambiguous. In both cases, it's the stories of these creatures that imprint in our memory.
An untraceable entity—digital cryptids have no origin story—they appear seemingly out of nowhere, growing out of the ever-evolving feedback loop of stories that surround them.
We were all unknowns on the early internet; our ‘real’ identities hidden and obscured through avatars and encrypted usernames written in leetspeak. Anonymity opened users up to new worlds and selves completely divorced from the IRL, allowing a certain type of freedom—from capital, censorship, identity. To reveal your AFK self was a huge faux pas, and would likely summon swarms of trolls—before Facebook’s real name policy and influencer culture dissolved the boundaries between the real and digital. Though a lot has changed—our digital footprint is tracked like a Reddit sleuth hunting down Bigfoot, identities split across multiple screens and devices as digital fragments—the user as cryptid can learn from these earlier models to avoid captcha—by market forces, algorithms, and the like.
A lot happens that we cannot see, and a lot exists beyond our human understanding. The weird provides a way of accessing these “out there” spaces past the limits of our ordinary reality. As Timothy Morton suggests in his 2016 book Dark Ecology: “In the term weird there flickers a dark pathway between causality and the aesthetic dimension, between doing and appearing, a pathway that dominant Western philosophy has blocked and suppressed.” Say goodbye to the fedora-tipping atheism of 2010s Reddit! We’ve come a long way since then—we haven’t done away with scientific hard truths, but as consensus reality evaporates, from the cracks emerge more weird possibilities, mysticisms, constructs, and conspiracies. There’s a sense that material and virtual reality has broken down to such an extent that everything is post-truth, nothing is real—as Erik Davis points out in High Weirdness: “[The] intertwining of reality and fantasy has become a crucial feature of what some scholars of religion now identify as ‘postmodern’ or ‘hyper-real religion’.”
We have never been modern, provokes the French philosopher Bruno Latour. Here the paradox of modernity is the division between “nature (the realm of objective, impersonal truth) and culture (the realm of subjective belief)”. But, at the same time, we constantly interact with hybrids between these groups. “This modern constitution allows the expanded proliferation of the hybrids whose existence, whose very possibility, it denies,” he writes. Scientific facts are built on a networked system determined by human categories, which tends to overlook the non-human—animals, plants, minerals, objects—whose existence is everywhere. There’s trillions of bacteria, fungi, and viruses that cohabit our human bodies, along with countless other inorganic materials: BPA, minerals, microplastics, volatile organic IKEA offgassing compounds—you get the idea. Surely we’re all cryptids to some degree when broken down into the bare parts.
The hybrid has been given many names across history—monsters, cyborgs, tricksters. In her A Cyborg Manifesto, Donna Haraway says that “we are all chimeras, theorised and fabricated hybrids of machine and organism; in short, we are cyborgs”. Myth-making has always played a part in our understanding of the relationship between human and machine, though social media has made this storyfication of both ourselves and the surrounding culture more potent, more personalised, more weird—as tailored as targeted ads. Our social media feeds form the crux of how we want to present ourselves to the outside world, which in turn shapes the reality we inhabit. We perform our digital realities, expertly crafting our personal lore, while our IRL selves lag awkwardly behind us. As Shumon Basar puts it: “Lore is the new myth you make about yourself. You live inside these myths mythically.” In a constant dialogue between the real and imagined, we are hybrid creatures, cryptids whose weirded reality is made up of layers upon layers of masks. I’m reminded of a meme depicting a stack of Wojak faces piled atop each other, with the caption: “I Have Buried Myself Under Too Many Layers Of Irony And Now I Can No Longer Remember Who I Really Am.”
Fictioning helps us navigate the unknowable—but it can also unlock new ways of seeing the world. In live-action roleplay (LARP), for example, the player channels the space between the real and imagined to explore new ways of being. “Bleed,” a term common in the LARP community, describes the lucid space that emerges when the two collapse in on each other—no longer a matter of fact or fiction, possible or impossible. For the Italian philosopher Federico Campagna, what we understand as reality depends on what society says is possible to think and imagine. He argues that “we must imagine a new set of reality principles that would allow for a new range of the possible to emerge.” The internet similarly allows us to experiment with reality constructs—as personal narratives morph and mutate, they form part of an eternal scroll of possible fictions through which the cryptid moves.
Situating ourselves in time and space in relation to the digital is becoming increasingly difficult and disorientating. The increased smoothness through which we move through online spaces has led to a flattening of the image—think algo-friendly trends like NPC influencers, whose uncanny online rhetoric reads like an AI text-to-image prompt, or viral face enhancing filters that standardise our features and make us resemble those strange dolls that populate the clear-pilled corners of the net. We’re beginning to think and speak like machines: our on-screen behaviour is primed for the algorithm, quantified by clicks and likes, monetised by engagement for smooth brains.
One particularly helpful way to understand this postmodern condition is through philosopher Vilem Flusser’s figure of the vampire squid or Vampyroteuthis Infernalis—as set out in the eponymous work of theory-fiction. This deep-sea celaphod, far enough from ordinary human existence to feel like something of a cryptid itself, evades conventional taxonomy—it’s suspended between the “dry” technology of the camera and the “wet” technologies of ink clouds, a dichotomy which parallels our own networked behaviour. As humans we’re made up of physical hardware—we fuel our bodies with food, take transport, move through the physical world. Online we are immaterial, disembodied, frictionless, and non-human. We tap and scroll uninterrupted between apps and browsers, while the algorithm locks us into the information flow—it’s inescapable, invisible, designed for human consumption, through censorship and extraction.
In Capital Is Dead, cultural theorist McKenzie Wark describes a new post-capitalist era defined not by goods, but vectors of information. The philosopher Byung-Chul Han agrees: “It’s not objects but information that rules the living world. We no longer inhabit heaven and earth, but the Cloud and Google Earth,” he writes. “The world is becoming progressively untouchable, foggy and ghostly.” Both relate to Benjamin Bratton’s theory of The Stack, which describes computing systems as a global megastructure that has overtaken all other kinds of governance.
The Stack is also a cryptid that “does and does not exist”—albeit a gigantic one made up of various kinds of computation, like smart grids, cloud computation, next-gen interfaces, addressing systems, generative AI, and chatbots. Through its intersecting and competing platforms—coding languages, frameworks, databases, frontend and backend tools, APIs—it dictates what we have access to, our experiences and desires. More than industrial infrastructure, how we travel through the stack informs how we experience ourselves and each other, both online and offline. It’s so big, so incomprehensible that it’s weird, like really weird—as sorcerous and concealed as a magician’s hat. But it’s also inescapable—it’s literally everywhere, from search browsers to Google Maps to Meta, an invisible mega-threat lurking in the shadows like Lovecraft’s Cthulhu.
Navigating through the shadowland isn’t easy—as we enter an era defined by The Great Weirding, there’s a pervasive sense of strangeness, synonymous with postindustrial techno-capitalism, which infiltrates every aspect of our lives, from computational systems to human-machine relations and social reality. This disorientation is all too predictable, and has become hyper-normalised—lest it sends us spiralling into the void. Since we can’t log off, we must keep moving or surrender—like the cryptid, we must morph and mutate in and out of fictions, stacking mask upon mask to reveal an identity, always lurking in the shadows, never fully there or out of reach.