A Human 'Non-AI Creative Art Model' Is Approaching You
In a world replete with A.I.-generated art, graphic designer Pablo Delcan had an idea for a revolutionary new tool: the “very first non-A.I. generative art model.”
“Send me a message with a prompt,” Delcan invited the public in a cheeky Instagram post in early December. “I will generate a drawing with black ink on paper and send it to you.”
That’s exactly what he’s been doing since. So far, he’s produced drawings from some 250 text prompts ranging in tone and length (“fog,” one reads simply). Many prompts are absurd: “A grim reaper frustrated with his computer.” “A T-Rex doing a handstand.” “A flower giving birth to a pencil.” Some are poetic (“a city made of clouds”) or personal. “I’m a little down,” one of the most recent submissions reads. “Can you draw a cute girl who says, ‘You can do it. Just get up and leave your warm blanket?’”
The whimsical work by Delcan’s Non-AI Generative Art Model, called Prompt-Brush, has delighted viewers, so much so that more than 2,700 prompts await Delcan in his online queue.
Delcan is originally from Spain and now lives in New York, where he designs books and book covers and contributes visuals to The New York Times. He’s busy putting paintbrush to paper as artists struggle to make sense of the implications of generative AI tools on their work—and on creativity itself.
While the project started as a joke on the social media platform, the designer earlier this month launched a dedicated Prompt-Brush 1.0 website where visitors can input a text prompt and email address and get a free drawing in return. They can browse a catalogue of drawings produced so far and scan prompts waiting for Delcan’s human touch. The site also features an amusing little tutorial explaining how olde-timey ink drawings like Delcan’s get made.
“A brush is used to draw by dipping it into black ink and then moving it across a piece of paper to make marks,” the description reads. “Light touches make thin lines, while pressing harder makes thick lines. It’s possible to make all sorts of drawings by connecting these lines.”
But while the project pokes fun at the ubiquity of prompt-driven creativity, Delcan said it’s less a comment on the debate over art and A.I. than something he does for the love of it. “I’m ultimately more interested in this as an experiment for myself,” he said in an interview. “All these drawings that I’ve been making for people, emailing each of them out to a single person, putting some thought into them. I think it’s been the most real exchange and connection I’ve ever experienced with my own work.”
After drawing an image, Delcan photographs it with his phone and increases the contrast on his computer for a clean, consistent black-and-white look.
As creative work aided by artificial intelligence proliferates, some artists are excited about the creative possibilities it affords, while others fear their work is getting scraped from the internet to train AI datasets without their permission, compensation or credit. Many creatives, Delcan included, hold both views simultaneously, though Delcan leans heavily toward enthusiasm. “The positives are many,” he said. “It’s the dream assistant that can help generate sketches and develop ideas. Maybe I’m just naive about it all,” he added, “but I think that giving anyone really powerful tools to be used creatively is a good thing. Another thought is that AI art generators are going to give more value to human-created images and human ideas and experiences.”
The designer draws an average of 15 says prompt-fueled images on a day, but sometimes as many as 50. They typically take a minute to create, two minutes for a more complex drawing and maybe five minutes if he draws a few versions.
Still, despite the speediness of the task, those who submit prompts to Prompt-Brush shouldn’t expect an immediate response. Unlike image-generating algorithms, Delcan the human art generator needs to take coffee breaks and catch some shut-eye every now and then.