A.I. Can Now Create Art Based Upon Text Prompts - DALL-E 2 and CrAIyon (Formerly DALL-E Mini)

"a painting of a fox sitting in a field at sunrise
in the style of Claude Monet”
Image by DALL-E 2, OpenAI.com

One thing people have often said is that machines will never replace creative artists because they lack the imagination and emotion required to create anything meaningful.

A.I. technology is now good enough to work as writing assistants, including sites like Jasper and Rytr, where you simply tell the A.I. what you want to write about and it'll produce a complete article on that subject. The article may not be perfect but it can sure save you a lot of time.

DALL-E 2: Realistic images and art from a description in natural language.

With the release of DALL-E 2 there is now an A.I. that can produce realistic (or artistically styled) images based entirely on text prompts in seconds. We're not just talking about collaging a few images together here either. DALL-E 2 can create images where everything looks fairly seamless and purposefully included.

Rather than me trying to describe it, watch the short video below to see how it works and some examples of the results.

As you can see DALL-E 2 could be incredibly useful for all kinds of applications and industries. Basically anyone that needs creative imagery could probably benefit. 

Note that, because it's so good, DALL-E 2's training has excluded things like images of famous people or horrific and NSFW pictures in an attempt to curb morally questionable output.

Many visual artists have commented that this technology puts them on notice in terms of being the next industry to have machines take over from humans but, as a commercial visual artist myself, I just see it as another tool. 

Imagine having an idea and seeing it visualized in seconds. For the average non visual artist that would be exciting but imagine what someone with imagination, vision, and creativity might do with this technology?

Unfortunately DALL-E 2 is not available for general use, although you can join a waiting list if you'd like to try it as soon as it becomes available.

CrAIyon (Formerly DALL-E Mini)

Although not nearly as refined as DALL-E 2 CrAIyon is the same basic principle, is free to use, and you can share images for personal use. If you want to use the images commercially you'll need to contact the developers.

Nine images generated by CrAIyon for
"David Bowie sitting on a star in space"
Check out 'Jake the Peg' Bowie (Image 4).

Unlike DALL-E 2 CrAIyon sources its images from the internet with few restrictions. For example you can enter celebrity names and you will see an attempt to create images of that celebrity. Though it seems to struggle creating recognizable faces of humans in general so expect some very wonky likenesses.

To be honest CrAIyon is not that great. I think it purposefully distorts most human faces based on my simple request to give me "a photographic image of David Bowie". You can clearly see the images are all drawn from classic Bowie photos but it's hard to believe an A.I., even of this level, would mess up the faces to the point of being demonic and zombie like. (Though just quietly I think Bowie would love these images as concept art for a future album).

"A photographic image of David Bowie".
This just screams of purposeful distortion
of human faces since it also does this to
non-celebrity human faces.

Despite this limitation CrAIyon produces some really fun images when you request specific art styles (or just request a painting, cartoon, or illustration) rather than general or realistic images.

Some of those cartoon images of David Bowie (below) would be really fun to turn into character animations or use as a style guide for a Bowie animated music video.

---o ---o--- o---

"A cartoon of David Bowie".
Some of these could be the inspiration
for an 'edgy' Bowie animated music video.

While CrAIyon is fun to mess around with it's no DALL-E 2. It's almost the equivalent of A.I. creating art with actual crayons in comparison.

Personally I'm very excited to see where this technology goes and hope that DALL-E 2 and similar future A.I.'s will be accessible to anyone.

Being able to see a finished idea for an image in seconds opens up so many possibilities for speeding up workflows and pushing the boundaries of the creative process. To me it's a very exciting future for visual artists to embrace rather than fear.