Make way, Picasso. A new neural network can draw images from text.

DALL E, a portmanteau of the names of artist Salvador Dalí and Pixar’s WALL E, can take any text and turn it into an image. The system uses a neural network trained on billions of images and text samples. It’s one of a growing number of AI projects that can mimic, but not replicate, the creative output of humans.

“Because natural language is constantly evolving and highly dependent on contextual nuances, teaching a machine to understand language well enough to draw a picture is a huge achievement,” Tamara Schwartz, a professor of cybersecurity at York College of Pennsylvania, said in an email interview. “Imagine a police officer, who is a rare talent, who can draw a picture based on a witness statement.”

Created by AI research company OpenAI, DALL-E works by collecting massive amounts of data from the internet. The data is then processed by a natural language model and trained to produce images from text. DALL-E works in a similar way to the recently released GPT-3, a language model created by OpenAI that can be forced to generate original text passages. Trained using half a trillion words of internet text, GPT-3 can produce surprisingly lifelike text.