AI created Terminator's BMW car

       This visual was created, after my prompt, by an AI art tool in 10 seconds. Is this real art? Or maybe a way for wannabe artists to make an impression? Fabricated garbage? And even worse, could this lead to the death of art as we know it?

Stable Diffusion is a text-to-image deep learning model, that in simple words generates detailed images conditioned on text descriptions. Its open-source release was announced in August 2022, which is quite huge, as content creators who want to get started with AI image generation now have an affordable option with three critical advantages over other solutions:

1. It’s open to developers to implement it in their apps without any oversight or censorship from the model maker. (we have already done so in Pobuca Experience Cloud – stay tuned to the next release in Q1 2023 )
2. It’s capable of producing images that other tools will not, due to copyright reasons.
3. You own the images you produce.

We could keep talking about the merits and the tech behind Generative AI and Stable Diffusion for quite a long, but is equally important to put the focus on how a new role is being born because of AI:

A new type of content creator will have the skills to describe texts to AI and apply such filters that a unique outcome will be created! Generative AI is a revolutionary and disruptive technology, that will produce a new generation of artists who embed AI in their workflow.

Now, let’s see step by step how this is done. 

We are going to use Playground, a user-friendly front-end for Stable Diffusion and our task is to prepare images for a campaign that is related to Ancient Greece.

I started with this prompt: “An ancient Greek playing a game and thinking”. To me, the result below is spot-on, this was

 exactly what I was envisioning in my mind’s eye when I gave the prompt!

Now, let’s say that we want a modern campaign, something that looks more realistic about Ancient Greece. The next prompt could be something like:

“Portrait photo of an Ancient Greek old warrior chief, tribal panther make up, blue on red, side profile, looking away, serious eyes, 50mm portrait photography, hard rim lighting photography”

Let’s also have his wife: “Ultrarealistic, ancient Greek old woman, portrait, cinematic lighting, award-winning photo, 80mm lens”

This story nicely illustrates an important rule of prompt engineering: the more specific the idea you have of what you want to see, the more relevant words you’ll need to add to the prompt to produce it. Now let’s use a celebrity for our main character:

“Keanu Reeves portrait photo of an ancient Greek old warrior chief, tribal panther make up, blue on red, side profile, looking away, serious eyes, 50mm portrait photography, hard rim lighting photography”

Maybe Keanu Reeves on a horse?

“Keanu Reeves portrait photo of an ancient Greek old warrior chief, on a horse, tribal panther make up, blue on red, side profile, looking away, serious eyes, 50mm portrait photography, hard rim lighting photography”

Let’s switch to Tom Cruise, the same prompt but with another celebrity:

“Tom Cruise portrait photo of an ancient Greek old warrior chief, on a horse, tribal panther make up, blue on red, side profile, looking away, serious eyes, 50mm portrait photography, hard rim lighting photography”

Or, maybe Robert DeNiro?

“Robert DeNiro portrait photo of an ancient Greek old warrior chief, tribal panther make up, blue on red, side profile, looking away, serious eyes, 50mm portrait photography, hard rim lighting photography”

Here’s a tip that I am using to build successful prompts: 

Imagine you’re writing a description of your desired image for a web page, which people will be putting into Google to search for your image. Ask yourself what terms users would probably type into Google to find your image if they had already seen it and wanted to locate it again. Make a list of those terms, and then use all of them in the prompt.

The applications of Generative AI and art tools are limitless. From personalized images in content marketing to virtual worlds and characters in gaming based on user input.

Maybe some artists will be so good that they will be hired by brands to create AI unique drawings to be offered to their loyal customers as NFTs – the sky is the limit!

Finally, maybe the visual artists of the future will be poets! Poets will be able to express themselves with unique prompts in a way that Generative AI will produce the finest art! Like they already do with lyrics in music: with their words they visualize melodies – now AI will visualize their words!