The Development of AI in Media & Content Creation

What if we told you that the image above was created entirely by artificial intelligence? The future of media and content creation is here – and it has the potential to revolutionise how we design, produce and market content for the foreseeable future.

It’s easy to be cynical about the authenticity of an AI that is capable of blowing your most masterful piece of artwork out of the water in less time than it takes to make a cup of coffee, especially in an industry already influenced so much by hyperbole. However that truly is, in a sentence, what programs like Midjourney are already capable of doing – and it’s incredibly simple to do.

The development of AI Art as we know it today first began around 2015 – only 7 years ago – where a class of software algorithm known as “Neural Style Transfer” was developed. Using complex looking code and a lot of computer processing power, the computer was able to process an image online, manipulate its data, and transform it into the visual style of a second image. For example, if you wanted the “Mona Lisa” painting in the style of Vincent Van Gogh’s “The Starry Night” (because why not?), the algorithm would transfer the visual style of “The Starry Night” and merge it into the image data of the “Mona Lisa” painting, creating a blend of the two.

 

This new art medium was developed further with the conception of a “Generative Adversarial Network” – a variation of 2 combined Neural Style Transfers that was now entirely classified as “machine learning” – in plain terms, this AI was purposely rogue. The difference now was that the AI could scour through thousands of online images, learn their intricacies, and recreate a completely unique version of all of the images it has just processed.

This new form of media stayed relatively niche for a while, with a few wider publicised exceptions including an incredible piece of physical art by Mario Klingemann titled “Memories of Passersby I” – an infinite real-time production of one-of-a-kind AI-generated paintings. These are not predetermined programmed combinations, they are created on the spot by a machine. 

 

 

Alongside further AI developments such as “CLIP”, an AI that analyses images and labels its data in written English, images across the internet were now being sorted into giant online databases with accurate tags. This birthed the creation of text-prompt based AI image generators, such as Midjourney.

 

Midjourney is one of a handful of revolutionary programs that creates an incredible image based off a single line of text. For example, if you were to type “Royal Skeleton” into Midjourney, this is what you would get in seconds:

 

 

Again, each image created is completely unique – even if you input the exact same prompt multiple times. In fact, Midjourney is built for the user to get as much accuracy out of their prompt as possible. It firstly generates 4 images from your prompt, each with a slightly different style or composition. You can then choose to receive more variations out of any of those 4 images, or you can upscale them for higher quality images. 

 

 

There are custom variables that you can stick at the end of your prompts for even more amazing customisation. For example, “–ar 21:9” will give you an image in a widescreen aspect ratio, “–style 60000” will tell the AI to go as crazy and as abstract as possible with your prompt, etc.

 

 

Now that we have established that this is no gimmick, what are some of the practical uses in our industry that we can get out of Midjourney right now?

 

Concept Art

This poses the most danger to aspiring Concept Artists – it really shines in producing amazing proof-of-concepts almost instantly, making iterations far more efficient, accessible, and cost-effective whilst not compromising on quality. This could be particularly useful for Movies, Video Games and Animation. It has the power to visualise your very own fantasy (or non-fantasy) world.

 

 

Creative Direction / Inspiration

Expanding upon Concept Art, Midjourney can be particularly useful for Creative Direction or Inspiration. For example, if you were struggling for ideas with designing a poster, in seconds you can create a concept to steer your ideas and align your focus. You might be struck by the shade of orange used, or inspired by the overall composition.

 

 

Logo Creation

Midjourney is not just a tool for abstract paintings – whatever you have in mind to create…you can create!

 

Texture Creation

 

“Photoshop” jobs with no technical skill required

Yes, this is Elon Musk as a dog.

 

NFTs

A reminder that every single image you create is one-of-a-kind. Midjourney has no restrictions on the usage of images in the Cryptocurrency market. In fact, by default you are the official copyright holder of every image you create!

 

 

Other Examples

 

• An unlimited supply of assets available to create
• Great for extremely fast turnaround times – the average duration for image creation in Midjourney is less than 30 seconds
• There is no limit on how creative you can get
• An array of styles – 2D, 3D, Clay, Pencil, Painting,Comic, Origami, etc. Any artystyle you want – just specify!
• The potential to go “Viral” on social media just got easier – just like the Elon Musk example above, you can hop on trends in seconds and create striking imagery

For further accreditation, large brands are jumping in on this too. Nike recently collaborated with a company to create the first ever shoe design that was NOT designed by a human.

 

 

Innovation is on a rapid trajectory, with the capabilities of AI going beyond the constraints of just imagery into Video Content and 3D Art as well. 

This is really just scratching the surface, and really the beginning of a whole new world, and I hope you are just as excited about the possibilities this opens up for the industry moving forward!

Author: Harry Fox