AI Produced These Imaginative Artworks By Taking Its Own Creative Liberties

By Mikelle Leow, 08 Apr 2022

Image via OpenAI

The time is here. Artificial intelligence is officially pretty talented at making art.

Last year, OpenAI, an AI lab funded by Microsoft, introduced DALL-E, a technology that can create impressive graphics, spanning numerous styles, with written descriptions as its only prompts.

Truly. You could tell DALL-E to produce illustrations of “a baby daikon radish in a tutu walking a dog” and it could do it with ease. The system’s moniker, by the way, combines the names of surrealist painter Salvador Dalí and Pixar’s WALL-E the robot.

DALL-E is then able to imagine other elements that might belong in its scenarios and “fill in the blanks.”

“Unlike a 3D rendering engine, whose inputs must be specified unambiguously and in full detail, DALL-E is often able to ‘fill in the blanks’ when the caption implies that the image must contain a certain detail that is not explicitly stated,” OpenAI explained last year.

But like many artists, DALL-E was still figuring things out. And at long last, it has now evolved into DALL-E 2.

Prompt: “A bowl of soup that is a portal to another dimension as digital art.” Image via OpenAI

DALL-E 2 can do the same things as its predecessor, but quicker and better. Artworks it produces are more higher-res and more detailed.

According to TechCrunch, DALL-E 2 stands out from the previous version in its use of a diffusion model, which counterintuitively makes graphics clearer by adding noise and gradually removing it. DALL-E 2 also fills in gaps by “inpainting,” conjuring up objects such as a non-existent bird or flower species—and it can do so by including the appropriate lighting and shadows.

The technology is so remarkable, OpenAI has its qualms about releasing DALL-E 2 to the public. The new program is thus accessible by invitation only; and yet, the firm remains cautious by banning obscene or extremist terms from being added into prompts.

Interested users will have to sign up to join the tool’s waitlist for a shot to be part of the program. The rest of us will just revel at the results:

Prompt: “An astronaut riding a horse in a photorealistic style”