She’s out there somewhere, lurking in a parallel universe of possibilities. All you have to do to summon her into being is type the right prompt into an AI image generator.

Like a digital incantation, the words will reveal the eerie face of a middle-aged woman with dead eyes, a vacant stare, and a disturbing grimace.

Her name is Loab (pronounced like ‘lobe’), and she was “discovered” by a Swedish-based artist who goes by the name of Supercomposite on Twitter.

Supercomposite is among the first wave of modern creators to explore the realms of text-to-image AI generators. This year, while experimenting with negative prompts (which ask machine learning algorithms to find the extreme opposite of something), the artist stumbled across a creepy face.

When Supercomposite ran the prompt again, they said the same woman came back, this time next to the word ‘loab’.

“The AI reproduced her more easily than most celebrities. Her presence is persistent, and she haunts every image she touches,” Supercomposite wrote on Twitter in a September 2022 thread about Loab’s discovery.

“Take a seat. This is a true horror story and veers sharply macabre.”

With a hook like that, it’s no wonder Loab proceeded to take the internet by storm. The image of this mysterious woman is now so well-known she even has her own Wikipedia page.

Part of Loab’s mystery is what she represents. Loab’s figure has become a sort of modern-day ‘tronie‘ – a type of artform from the Dutch Golden Age that exaggerates the expression of a face – one that doesn’t represent a person but an idea.

The allegory of Loab is just a bit more terrifying than, say, the subject of the more famous tronie titled Girl With A Pearl Earring. More profoundly, it wasn’t made by a human artist who can tell us more about the idea they were trying to represent.

Among the hundreds of Loab iterations Supercomposite has summoned into being, many feature dismembered or screaming children in the background. Some AI-generated images were so grotesque that the artist decided not to share them publicly.

“I was ripping Loab apart and putting her back together. She is an emergent island in the latent space that we don’t know how to locate with text queries,” writes the artist on Twitter.

“She finds everyone sooner or later. You just have to know where to look,” Supercomposite adds.

Loab has captured the world’s attention for more than just her nightmarish qualities. Plucked from the abyss through what Supercomposite calls an “emergent statistical accident”, the eerie woman represents a new era of creativity we may or may not be ready for.

Brendan Murphy, a photographer and lecturer in digital media at Central Queensland University in Australia, spends much of his free time thinking about the future of AI and sampling image and text generators.

With the technology recently exploding, he thinks the art world is headed for a paradigm shift, much like when photography arrived on the scene in the early 1800s.

Today, when Murphy uses AI to make art, he thinks of it like landscape photography, wandering around a place and looking for interesting things to capture. Except, in this case, the landscape he’s exploring is a sort of parallel universe of human art.

After all, AI generators are trained on human knowledge, culture, and traditions of art, which means that we could have plausibly done anything they create.

These unrealized possibilities are now out there for people to find, and Murphy and Supercomposite are among the first to join the hunt.

“There are things that you see that interest you, that you really want to amplify, and really want to go in that direction,” Murphy explains to ScienceAlert.

“There’s no reason to go on these paths. And there are probably really good reasons why people have never gone down these paths. Because it’s probably never going to impress anyone or sell anything.”

That doesn’t mean using AI to make art is frivolous. Instead, Murphy says AI is a tool artists can use to further their artistic practices. And every once in a while, a precious figure like Loab emerges from the abyss.

“I think the thing about Loab is, it is a great story. It’s not just the technology. It’s looking at what drives the technology. It’s looking at the possibility of the technology,” he explains.

“And I think that is great. I think that is a valid artwork. Much more valid than just making a particular AI image. There’s a lot of thought, a lot of experimentation, a lot of iterations.”

Anne Ploin, a digital sociologist at the Oxford Internet Institute researching the potential impact of machine learning on creative work, shares a similar view.

“AI models can extrapolate in unexpected ways [and] draw attention to an entirely unrecognized factor in a certain style of painting,” Ploin says.

“But machine learning models aren’t autonomous. They aren’t going to create new artistic movements on their own.”

Murphy and other art experts think it’s doubtful that AI will erase human creativity, at least not entirely. Art, after all, only exists if humans value it, and as a species, we tend to be pretty biased about our abilities.

Going forward, AI-generated artworks can prompt us to question artistic traditions and explore our emotional reactions to images, Murphy says.

But we are entering a world where many writing and painting services could conceivably become redundant, erasing the jobs of many ghostwriters, illustrators, designers, and photographers.

The explosion of AI-generated art in recent months has sparked concerns that algorithms are ripping off artists by replicating their distinct styles.

The best human artists will no doubt continue to compete with AI, and Murphy suspects it will be those creators that lean into their human-ness that have the most success in the future.

A public face and identity that is tangible and real could be more crucial for artists than ever before. “Regardless of how apps like Lensa shake up the way art is made, the artist’s personality remains an important context for their work,” Murphy wrote recently in The Conversation.

ChatGPT is an AI-based text generator that was publicly released in November 2022. Its name stands for Chat Generative Pre-trained Transformer, with an engine based on assessing the probability of certain words following one another in a block of text, not unlike a supercharged version of your phone’s predictive text feature.

Like Loab, there appears to be a ghostly presence in the program’s responses that test the boundaries of human knowledge and creativity. Though it doesn’t take much for the veneer to fall away, exposing its wiring as a dice-rolling “parrot” that gambles on producing the right strings of words to fool us.

Any wisdom it provides is still left up to us to pick apart and judge. When asked how AI is changing art, ChatGPT agrees with Murphy:

” Artificial intelligence is changing the way that art is created, perceived, and experienced. AI algorithms can be used to generate novel forms of art, such as music, poetry, and visual art. These algorithms can also be used to analyze and interpret existing artworks in new and interesting ways. Additionally, AI technology is being used to create interactive art installations and performances that can respond to and engage with viewers in real-time. Overall, AI is making it possible for artists to create and explore new forms of expression, and for audiences to experience art in new and exciting ways.”

The boundaries of artistic practice are expanding, and Loab is just the beginning.