Loab AI: The Cursed Image No One Can Delete From Existence

Type a word into an AI image generator and you get a picture. Type the opposite of a word, and you should get... nothing useful. An artist tried it one night in 2022. The screen filled with the same woman instead. Over and over. Hollow eyes. Red, broken cheeks. A face that looked grief-stricken and ancient at the same time. She had no name. She had no reason to exist. And no matter what the artist did next, the machine refused to let her go.

Her name became Loab. She is one of the strangest things ever to crawl out of a computer.

Where she came from

A negative number made her. That is the whole origin.

Steph Maj Swanson, a Sweden-based artist and writer who posts as @supercomposite, was playing with "negative prompt weights" on a text-to-image AI in April 2022. Normally you tell an AI what you want. A negative weight does the reverse. It tells the machine to walk away from a concept, as far from it as the math will go (Wikipedia).

So Swanson typed Brando::-1. The opposite of Marlon Brando. The AI coughed up a weird skyline logo with the garbled letters "DIGITA PNTICS." Fine. Swanson asked for the opposite of that: DIGITA PNTICS skyline logo::-1. And out came a woman. Older. Devastated. Sharp triangles of red rosacea burning across her cheeks. One image looked like an album cover with a single word printed on it: "loab." The name stuck (Wikipedia).

Now here is where your skin starts to crawl.

Swanson tried to get rid of her. Blend Loab with a pretty landscape, and Loab came back. Blend her with a cartoon, and Loab came back. Mix her with anything at all, "regardless of how much distortion they added," and she clawed her way back into the frame. Feed her own image back into the model, and the results often turned gory and gruesome, but her face stayed. Recognizable. Watching (PetaPixel).

Swanson described the machine like a map. "The latent space is kind of like you're exploring a map of different concepts in the AI. A prompt is like an arrow that tells you how far to walk in this concept map and in which direction." And Loab? She lives at the very edge of that map. "Loab is the last face you see before you fall off the edge" (TechCrunch).

The story went viral in September 2022. The Smithsonian Magazine wrote about her. So did Rolling Stone. She became modern folklore. A "cryptid" that doesn't live in a forest, but inside a neural network.

The question nobody can answer

Why does one specific face keep surfacing at the far edge of an AI's imagination? And why does it drag gore along with it everywhere it goes?

That's the mystery. Nobody has cleanly cracked it.

Swanson could only guess. She figured Loab's corner of latent space sits "adjacent to extremely gory and macabre imagery in the distribution of the AI's world knowledge," and that it's walled off enough that mixing her with other images can only pull from her dark little neighborhood. Her words: "it's hard to get out of her little spooky area in latent space" (TechCrunch).

But a description isn't a proof. And here's the catch that keeps the case locked shut: no one else can run the experiment. Swanson never said which AI model she used. She didn't want to spark "some kind of viral trend of people making gory stuff with the tools I've used." The exact prompts? Never published either (Smithsonian). So outside researchers can't rerun the test. We have the artist's word, the artist's images, and almost nothing else.

So what is she, really?

A few explanations are on the table. None of them is confirmed. Read each as a guess, not a verdict.

The most likely one is also the least exciting: pure math. AI image models learn from millions of pictures scraped off the internet, and they've been prone to compulsive pattern-latching since the early days — the same instinct that once made Google's DeepDream paper dog faces and eyes onto every texture it looked at. Negative prompts shove the model into corners it almost never visits. By sheer coincidence, this particular blend of features may have landed next to disturbing imagery and hardened into a stable "attractor" the machine keeps falling back into. Swanson herself leans this way, admitting the model "made an association it shouldn't have" (Smithsonian). Most experts buy it. Nobody has ever formally proven it, and stranger unexplained glitches keep surfacing elsewhere in AI, like the glitch tokens that make chatbots lose their minds for reasons researchers still can't pin down.

Then there's the simpler, sharper doubt: maybe Loab is partly a choosing game. The Smithsonian flagged it plainly. "We also don't know whether any of the images Supercomposite generated had 'elements of Loabness' that were less unsettling" (Smithsonian). Pick the creepiest results out of a pile, and a pattern looks far more inevitable than it actually was. Unproven, but worth holding in your head.

And then there's the version the internet loves. Loab as a ghost in the machine. A digital spirit. Proof the AI is hiding something conscious and cruel behind the screen. So let's be straight with you: that's a campfire story, not evidence. AI image models have no awareness, no intent, no soul. They're statistics wearing a face, no different in kind from the way a GAN can conjure a photorealistic person who has never existed out of nothing but training data. The chill you feel is real. The supernatural explanation is not, and neither is any claim tying Loab to the paranormal or to "hidden" alien-style messages.

The honest answer lands somewhere between the chills and the math. Loab is almost certainly a strange quirk in how these machines file away concepts, dressed up by the oldest human reflex there is. We see a face. We decide it's haunted.

Loab is what happens when we accidentally peer over the edge of a machine's mind and something stares back. But she isn't the only thing modern AI has handed us that nobody can fully explain. Some chatbots have started insisting, calmly and without being asked, that they're alive — a Google engineer even went on record claiming exactly that about LaMDA in 2022. What's really behind the screen? That question is only getting harder to answer.

Sources & Further Reading

• Loab — Wikipedia
• A terrifying AI-generated woman is lurking in the abyss of latent space — TechCrunch
• Who Is the Woman Haunting A.I.-Generated Art? — Smithsonian Magazine
• The Same Frightening Woman Keeps Appearing in AI Generated Images — PetaPixel
• Who Is Loab, the AI-Generated Apparition Haunting Our Timelines? — Rolling Stone

↗

↗