Artificial Intelligence has always had a place in the science fiction world. But now, it seems AI is moving from the sci-fi realm into our lives at a fast pace.
It really isn’t that big of a stretch to think that AI could one day replace designers or even copywriters. And all the while people do not realise it’s a computer generating images or text. In fact, most of this intro was written by AI!
DALL-E
DALL-E 2 is the most hyped AI for image generation at the moment, but access is limited to those fortunate enough to be granted an account.
If you’ve never used AI before, it’s really simple, just think of something you want and type it as though you asked someone on Fiverr to create you some artwork, such as:
“An astronaut riding a horse in a photorealistic style”
And you’ll end up with this:
The above is taken from the DALL-E website as an example of what their system can generate. They also have examples of “Teddy bears mixing sparkling chemicals as mad scientists in a steampunk style” and “A bowl of soup that looks like a monster knitted out of wool”
Every time you make a request it generates 10 different images for you to choose from, type the same thing again and you’ll get 10 different images again.
Youtuber Marques Browlee gained 24-hour access to have a play and created some interesting images, you can check it out here.
Craiyon.com
If you’re like me and don’t have access to DALL-E 2 yet (or ever) then you could always give Craiyon a go. This uses a ‘light’ version of DALL-E to generate images that do work, just not as well as DALL-E 2.
The interface is simple, type your text and generate.
Requests can take up to 2 minutes and can generate some…interesting results…
“The queen riding a corgi”
It doesn’t handle faces/animals very well and probably shouldn’t be used to generate art, unless you’re going for a cut-up magazine scrapbook style.
Midjourney
Mid journey is just that, a kind of midpoint between DALLE-2 and Craiyon. It handles requests much better and generates some pretty cool imagery as long as you’re not looking for photorealistic faces.
It’s limited to 25 requests and works through a Discord channel, the added benefit of this is that you can see what other people have requested and save their images too. This can lead to inspiration or some disturbing results!
I like retro-futuristic art, so I made some requests…
“Futuristic lithograph print planets”
“Retro futuristic space war”
It’s amazing to think a computer has taken those words, applied context to them and then generated these images.
Here’s a selection of some of the more odd looking results…
What will this mean for designers?
It’s by no means perfect at this stage, so anyone wanting pixel-perfect artwork/designs is still going to have to seek out a designer.
However, it’s a great tool to get a starting point either from a client’s perspective or designer. Sometimes the most challenging part of the design is understanding what the client wants. If they type in literally what they want it could reduce that initial time by giving designers a rough idea of where to go.
I can also see it being used to develop concept art for games, films, books etc. It can come up with some pretty interesting results that may have never been considered before.
It might be fun to watch. As designers, we all work on computers, sketching, painting and making human decisions quickly in an intuitive way with guesswork that seems impossible to replicate. I would love to see an AI try switching places with me for a day. How well could it do?
P.S That last paragraph was generated by AI, could you tell?
If you need more help with anything design or marketing related for your business, contact us at Colewood to see how we could help.
