2023 started off with the release of several different AI programs that can generate images from a text prompt that were, for the first time, readily available and accessible to the general public — Stable Diffusion, Midjourney, and Dall-E Mini/Craiyon were the three big releases. They are incredibly powerful tools, and can produce amazing images when used properly.
This leads many to ask: If anyone is able to produce high-quality pieces of art with no artistic abilities whatsoever, then what does this mean for the future of artists and designers?
To some, AI image generators are the greatest invention since the printing press that’s unlocking all kinds of new creative possibilities. For others, they’re tantamount to theft and a threat to anyone with a job in the creative sector. But the fact that can’t be ignored is that whether it’s seen as a godsend or sacrilege, there’s no going back. And anyone who works in image creation needs to know about them.
In this article, I’m going to explore what impact the AI technology behind the image generators that are now widely available to the masses is having in the creative industries, and what it might mean for the future.
The Artist in the Machine
The beginning of the year has seen an explosion of AI-generated images circulating around the web demonstrating the amazing capabilities of these programs that can seemingly create an image of anything you can dream up. They are incredibly powerful tools and can produce astonishing results.
From a dinosaur caught on a night vision trail cam to Darth Vader Fly Fishing in Montana to a futuristic McDonald’s in the year 2050.
Now apart from being a great tool for creating meme content, as you can imagine, these AI image generators have the ability to create images in not only any scenario but in any form.
You can depict your subject as a futuristic cybernetic robot, a stained glass window, or a Christmas tree ornament, as well as made out of any material – ice/wood/stone/marshmallows/sushi.
These too, are all in good fun and if you have a vivid imagination and a lot of time, you can keep yourself entertained for hours.
But what has really sparked a huge backlash against the technology and has people questioning if this is going to make artists and designers redundant, is the ability to create an image in any style of art. From oil painting, pencil drawing,d anime, to impressionism, spray-painted graffiti, and Japanese Woodblock.
If you have a favorite artist or artistic period, you can create an image in that style.
You only have to look up AI image generator illustrations to see some mind-blowing images or try one yourself.
How Does an AI Image Generator Work?
First things first – to understand how AI image generators will influence the future of art and design, you have to understand how they work. It’s only natural to fear the unknown, so let’s get to know them!
To create an image from a text prompt, AI image generators use two neural networks, also known as GANs (Generative Adversarial Networks). These neural networks work in tandem to produce the most accurate possible result from the program to the best of their abilities.
One neural network (the Generator) has been trained to analyze the text, retrieve the concepts, and recreate them from the prompt input by the user (that’s you). The other neural network (the Discriminator) has been trained to detect real (human-created) photos and illustrations and differentiate them from the AI-created images.
It then rates the results of the Generator to determine whether the created images look realistic enough according to what it deems ‘real’.
Once an assessment of the image accuracy is complete, if it fails to fool the Discriminator into recognizing it as a ‘real’ image, the data is sent back to the Generator. That system then learns from the feedback and sends back an altered image for further rating until the AI-generated image matches the control/template image.
Why this matters will start to become clear as we go on. Having a basic understanding of how an image generator works will provide a more accurate perception of the situation and how AI is likely to affect the art and design industries.
Will AI Image Generators Replace Designers and Artists?
We live in a society at a time when technology is advancing on a daily basis. This is nothing new to us these days. Artificial Intelligence has also been around for quite some time.
How long? … hmm, let me check.
“Hey Siri…”.
Now ask anyone who had a blockbuster franchise if advancing technology is the end of days, and they’ll probably tell you ‘yes’. But the reality is, although AI image generators will definitely shake things up and affect a lot of changes, they will not completely replace human-created art or design, and here’s why…
The AI generators are relying on two very important external factors to create images.
1. How the program was developed
Remember how we were talking about how the whole system works and how the Discriminator decides whether it’s an accurate enough image? It bases its accuracy on what it’s been taught and the millions of images it’s been fed for reference.
There are only so many photos and images out there and things that are uncommon, not often focused on, or have a lot of complex parts leave a lot of blanks for the programs to fill in which results in images that are inaccurate.
Take human hands, as just one example.
All of the AI image-generating programs have real issues with creating hands, almost always churning out images of missing or multiple, deformed fingers. There are online threads and memes solely dedicated to the amusingly inaccurate images of hands that come out of AI image generators.
On June 22, 2022, Twitter user @pikcatl posted the earliest viral meme about AIs’ poor ability to render realistic-looking human hands. The image, generated with DALL-E Mini, gained over 820 retweets and 8,000 likes in seven months (on the left, below). I decided to give it a go (February 2, 2023) with the same program to see what would happen now.
Since that first photo was made into a meme, Dall-E Mini, renamed Craiyon (and discussed here), which is now hosted on its own site, has undergone loads of updates and is constantly being improved. But even now, though the results are maybe a little better than the first renderings in its early stages, you can see (on the right) there’s still a looong way to go before it churns out anything even remotely realistic looking.
2. The User
These tools only really work if you know how to prompt them correctly. You may have seen some amazing images that have been created using AI image generators.
Take the slam-dunking Basketball Llama, for example. This image was created by Joy Zhang with the use of the closed beta version of DALL-E 2 AI image generator.
This is an epic photo. And if anyone can create a photo like this, by typing in “llama playing basketball”, we’re all in trouble. But we’re not. As Joy points out in her article: I spent $15 in DALL·E 2 credits creating this AI image, and here’s what I learned, it’s a little more complicated than that.
These are the images she started out with when she typed “llama playing basketball” into the prompt window.
Unless you know how to use the technology, it’s not going to give you the results you’re looking for. So no, not just anyone can create amazing works of art, incredible photos, or cool graphics.
The program is only as good as it’s been developed and how well the user knows how to work it. Although there are a lot of stunning images out there created by AI image generators, these image generators are not without their limitations.
You also have to remember that most of these images have been created by people just playing around with the programs because it’s all relatively new, and they have a lot of time on their hands.
It’s like those Trick Shot videos you watch of someone throwing a piece of bread over their shoulder without looking and it lands in the toaster. They are only showing you that 5-second clip…. that took them two weeks to get and where you don’t see the other 238 tries that were missed.
And to put it all into perspective, the people who created those amazing images using AI image generators also probably didn’t have a client with a deadline asking for something super specific that’s also keeping in line with their brand identity.
The people who created those amazing images using AI image generators also probably didn’t have a client with a deadline asking for something super specific that’s also keeping in line with their brand identity.
This is Not the First Time
This is not the first time that advancing technology has threatened to put an end to the livelihoods of those working in the creative/artistic communities.
When the camera was invented, people thought that it would replace human artists. It didn’t, and photography flourished into its own art form, creating a whole new industry.
Photography did, however, put a big dent in the business of portrait artists. If you are working in the creative community, you are going to want to make sure that you are not ‘the portrait artist’ of this time.
A more recent example was the introduction of programs like Photoshop and other image editing software.
When they came along, people were questioning whether professional photographers would become unnecessary. If anyone can take a crappy photo and edit it to look good, why would you need to pay a professional photographer? That, of course, didn’t happen either. And now image editing software has been embraced by photographers and is routinely used to enhance their photos.
Everyone knows that almost all of the images in the fashion and advertising industry are tweaked nowadays, whether it’s to make teeth look whiter or the model look skinnier.
Back in the day before image editing software was the norm, people were a lot more forgiving under the notion that you get what you get. But now it’s expected that a professional photographer should know how to edit photographs. Any photographer that did not keep up with the advancing technology of the time would be hard-pressed to find work these days.
Are Illustrators the “Block Buster Franchise Owners/Portrait Artists” of This Era?
Now arguably, illustrators have a little more to be concerned about when it comes to this technology. Painters can maybe rely a bit more on their personal style rather than the subject of the painting as their selling point. Graphic designers can rest a little easier knowing that computers don’t understand that form has to follow function, and that generic output is not what most clients are looking for.
Photographers need not panic because even if the tech advances to the point where it can actually recreate human hands properly, there will always be a need for photographers to capture events to record important moments in history. I’m also pretty sure no one will ever want a baby or wedding album full of AI photos.
No one will ever want a baby or wedding album full of AI photos.
Illustrators, on the other hand, are going to have a tougher time. Illustrators usually have to illustrate what their clients want according to their specs. If your client can get what they want from an AI image generator, why would they hire you?
If you’re an illustrator, this is why it’s important to know about AI Image generators. You need to know if these programs can do what you do.
There are various different forms of illustration from medical, manual, storybook, and textbook illustration to cartoon, fashion, and digital illustration for print, gaming, and other online use. The AI image generators are better at some types of illustration than others and they are not without their limitations.
That said, right now, even in its infancy the tech for AI-generated illustrations has produced some pretty amazing results. All of these programs are right now in their testing stage – they’ve been released to the public because the developers know people will let them know EXACTLY what’s wrong with them (and they have!), so they know where they need to focus their energy on making them more accurate.
Although you won’t be out of a job tomorrow, it’s just a matter of time before AI Image generators are able to create illustrations that are good enough that people won’t need to hire an illustrator.
It’s easy to flip the bird and dismiss AI technology as wrong and the destroyer of the creative process, especially when it seems like it’s coming for you. It’s not. It’s just technology evolving. Though having seen some of the memes out there, it seems like a lot of people are taking it personally.
There’s a lot of anger there and it’s not going to do you any good. Rather than joining those who have had a knee-jerk reaction of rebelling against AI image generators, you need to be on the smart end of this transitional period we’re all going through where AI is here.
We either have to get with the program or be that jaded, ornery person sitting on the front porch yelling: “You know in my day, there used to be a thing called creativity and we used to draw things with a pencil!” at random strangers passing by.
Here’s the good news… with change comes new opportunities.
The creative decision-making part of the design process is always going to be better when it is undertaken by someone who has a natural talent for or has been trained in the creative fields.
This is your opportunity to climb aboard the AI train to be part of the evolution.
Opportunities Created
There will be a lot of things that AI-generated images will be able to offer in all aspects of the creative realm that were not available before. That doesn’t mean that you, as a graphic designer or illustrator or photographer or artist, or anyone else in the creative domain will become obsolete. It just means that those working in the creative fields who are in the know about how to use this new technology and are able to incorporate it into their process will be those who are in demand.
Prompt Engineering is the new buzzword for what is soon to be a highly sought-after job skill.
Prompt Engineering is the new buzzword for what is soon to be highly sought after job skill.
Use AI as a Tool to help you
If you look at AI image generators as a tool to help you, as opposed to your competition, you will come out on the winning end. Remember: it’s called AI because it’s Artificial. It doesn’t have feelings, it can’t understand the passion of your client behind their business and it’s definitely not hell-bent on putting you out of work.
Once you remove the idea of it being the enemy, you will open up a whole new realm of possibilities and opportunities. Here are a few ways you can use these AI image generators to your advantage.
Brainstorming
Getting the creative ball rolling can sometimes be a bit difficult. We’ve all been there. Using the AI image generators to input your project goal just to see what it spits out is a great way to ignite a creative spark and use it as a starting point.
Preliminary Mockups
When you get a job to produce something for a client, it’s usually a process of trying out different ideas to see what resonates with them, followed by revisions and tweaking to produce the end product. Instead of spending a few days sketching up or putting together different concepts, you could use the AI image generator to quickly produce rough ideas to present to your client.
Collaboration
Incorporating AI-produced elements into your design is fair game as there are no rules and nothing is copyrighted. Things may change in terms of what you are legally allowed to use commercially, but right now take advantage of the free-for-all.
Take Away
AI is here and it’s going to be changing the way things are being done. You can either get on board and embrace it and find out how to use it as part of your process, or you can dismiss it entirely on principle and get left behind.
The more you know about the advancements in AI image generation, the more likely you are to be able to use it to your advantage and cash in on the new opportunities and jobs that it will open up for creative people.
