The AI Imagery Buzz

The AI Imagery Buzz
Date:

29th Jul 2022

by:

Breon Snowdon

Hey, want to see a Giraffe with arms punching an alien? Of course you do and now you can thanks to the latest internet trend. As predicted in our upcoming trends guide at the start of 2022, images created by Artificial Intelligence have exploded into the public consciousness. But are they a passing fad, a useful application of technology or a genuine threat to creativity.

I first watched a short documentary in 2016 about how a new ‘original’ Rembrandt painting was created 347 years after his death, using some clever code, intelligent algorithm and a dataset of Rembrandt’s back catalogue. The results are quite incredible, but the whole project was based around one outcome… jump forward to 2022 and AI imagery has leaped to the next level. With the recent unveiling of DALL-E the GAN scene has just exploded. 

AI imagery has been around for a long time but in the last few months has exploded with ‘prompt’ generated services, generative adversarial network (GAN).

So what is AI generated imagery, an AI model that can generate images from any ‘prompt’ you input. The model is trained by looking at millions of images from the internet with their associated captions. Over time, it learns how to draw an image from a text prompt.

Some of the concepts are learnt from memory as it may have seen similar images. However, it can also learn how to create unique images that don’t exist such as “the statue of liberty  landing on the moon” by combining multiple concepts together.

You can try it here; https://www.craiyon.com/ , just type in a ‘prompt’ something like; “ninja cat” or “godzilla baking a cake illustration” which will then present some interesting results. 

Very quickly other services have launched, along with algorithm updates with much better quality and more variations but now with monetary options too. With all the attention we signed up to MidJourney a discord based service that was showing some amazing results.

news_202207_01.jpg

(Images Credit: MidJourney users)

Now the results are incredible BUT it opens up a very emotive and interesting discussion about AI created art. These algorithms are learning off data online; learning different artistic styles and techniques, and with every ‘prompt’ accelerating its machine learning. So where does this leave the original creators of these styles or current artists trying to create new art? And are the tech companies putting the revenue they get back into the creatives?

news_202207_02.png

(Image Credit & Prompt: cute tentacle monster by realdesign via MidJourney)

I am a massive advocate of tech in art, we still use a custom randomiser script we wrote years ago to create some of our bespoke designs and elements, so I relish innovation in the tech but I worry about the damage to the creative process.

So my question to YOU is; Is this technology a good or bad evolution for the creative industry? It’s here and how we work with the medium will dictate its future. What do you think?


REF
DALL-E 2
Generative Adversarial Network
Google Imagen
A Thread: Dall-E or not Dall-E
41 Creative Tools to Generate AI Art
A Twitter Thread