Will AI fashion replace fashion design software or a fashion designer?

Will AI fashion replace fashion designers?

Jump to video version of this blog.

I’d been waiting for this question, and I finally got it the other day on one of my videos:

How do you feel about AI and how it will make a designer’s job obsolete?

And my first instinctual answer was, “Will it?”

Another really great digital fashion designer, The Fashion Tech, has been playing around with Midjourney, and she’s been posting some really cool pictures on Instagram.

The other day, she posted a flat sketch, that was generated by Midjourney, and I was floored!

I had been saying I was going to try Midjourney for weeks, but seeing that kind of lit a fire under me and had me thinking, “OK, I need to see what’s good!”

Now, there’s a lot that AI can do, and it’s being used in many areas of the fashion industry already. And I think because we’re now seeing more AI art generating programs, free or inexpensive programs at that, there’s a lot of concern around whether AI will start replacing designers.

First, let me show you how Midjourney works, and then I’ll give you my opinion about these AI programs.

Using AI app Midjourney

The first thing to note is that you need to join Discord and then add the Midjourney server.

To do this, you’d just need to go to Midjourney.com and then choose “Join the Beta”. From there, just follow the prompts to sign up for discord and add the server.

Go to Midjourney.com to sign up and join the Discord server.

This is where it got a little confusing for me because I’ve never used Discord. When I first looked up how to start working in the program, I was told to click the Midjourney icon in the left menu, and then, scroll to any of the newbie rooms.

And you can do that. The only thing I don’t love about working this way is when you’re creating anything in those rooms, it can be hard to find your art because while it’s generating, others are also creating art and so your art starts scrolling and can easily get lost within everyone else’s.

So instead, I click the Discord Direct Messages icon at the top left, and I create my art there. Although you can’t see what others are creating, which can be helpful when you’re trying to generate different types of art and figure out what prompts to use (which I’ll talk about in a second), it’s a lot easier to keep track of your work.

Click the Discord Direct Messages icon and create your art there.

To create any artwork in Midjourney (or any art AI program), you need to describe what it is you want it to generate. The descriptions are called prompts.

With the newest version of Midjourney, which at the time of this post is version 5 (and can only be accessed with a paid subscription), you can type your prompts in full sentences. If you’re still on version 4, you’ll need to use keywords.

If you’re new to this, you’ll probably need a few tries to get what you want. And even if you’re not new, you’ll probably still need at least a couple of tries to get this exactly how you want. But let me show you an example:

Whenever you’re starting something new, you’ll start with the phrase “/imagine”, and hit return. Next, Midjourney shows a box that says prompt and an area for you to type your description.

I’m going to type “full body fashion editorial photograph of a black female model in Valentino inspired dress looking away from the camera, insanely detailed hypermaximalist hyper realistic, Unreal engine, dof, studio lighting”.

Start with the word "imagine" and then type your prompt description.

What can make or break the image are the details you provide to the AI programs. So I added full body because I want to see a full body picture, I want it to look editorial, I want her to be a black female model clothed in a dress inspired by Valentino, and I want her to be looking away from the camera. The last few prompts are phrases to improve the quality of the rendering. DOF stands for depth of field, by the way, which is a photography/videography phrase that means the distance between the nearest and furthest elements in a scene that appear to be "acceptably sharp" in an image. The last prompt “ar 2:3” is the aspect ratio.

Hit enter, and wait a minute or two. This is what I got the first time I tried this:

These are the pictures I generated the first time I used this prompt.

Let’s see what it gives me this time.

Now, once it’s done, it’ll give you 4 options to choose from, and you can decide to either upscale one of the four options and render it in a larger more detailed format, which are the U options. Or create more variations of one of the pictures, which are the V options.

Use the "U" or "V" options to upscale or create more variations of a picture.

Once you find an image you like, click one of the “U” options, once it renders click “Web” which will open the image in a new window. Then, click the save icon to save the image to your desktop.

Save the final image to your desktop.

Using Midjourney AI app for flat sketches and illustrations

I also tried this with flat sketches and fashion illustrations.

For the flat sketch, I typed "fashion flat sketch, Women’s Moncler inspired ski jacket, 2 hand pockets, 1 chest pocket, 1 back pocket, hooded, technical drawing, text callouts", and this was the result.

The result of my flat sketch prompt

I used the same prompts to create an illustration but changed “fashion flat sketch” to “fashion illustration” and removed technical drawing. And here was the result of that. (Note: this was was attempt; I didn't like the 1st iterations.)

The result of my fashion illustration prompt

Now as cool as this is, I will tell you that it may take a few tries for you to get what you want. I wasn’t expecting those prompts to be so important, but I mean, you’re just describing what you want, and the program is trying to interpret an image for you based on that description.

Using this approach, you can see how Midjourney can easily give you lots of options, but all aren’t always so great.

Also, when it comes to the flat sketches and illustrations, those are cool and you might be able to use them as a starting point, but the flat sketch may not be exactly what you want, and if you notice NONE of those illustrations match the flat sketch I previously generated so if you were thinking you could generate a illustration and a flat for it, that’s not how this works. You’d still have to draw one or the other.

So, how do I feel about Midjourney and AI fashion apps?

I think they're cool. Do I believe it’s going to make a designer’s job obsolete? Not likely, and definitely not any time soon.

The first thing I’ll say is that the fashion industry is very slow to adopt new technology. And although the pandemic sped up some things, I’m already hearing people trying to go back to “the way we used to work”.

AI is improving rapidly, but it still lacks the creativity, intuition and emotion that is necessary to create a fashion collection.

Fashion is very subjective, and although AI can generate lots of fun interpretations of a design, it’s based on data and an algorithm and can’t replicate the unique perspective and personal touch of a human designer.

And, if you’re creating fashion for physical garments, there are a lot of other skills like garment construction, material selection, sourcing, and production management that require the expertise and experience of a human.

Think of AI like any other fashion technology. Use it as a tool to help you streamline your workflow and free you up to do what we as human designers do best:

BE CREATIVE!

I saw a post recently on LinkedIn that I think sums this up perfectly. It said,

AI will not replace designers. Humans who know how to use AI will replace designers.