Will AI Replace Graphic Designers? | Studio Maya
Written by
Creative Director

Getting straight to the point: no, you’re not going to be replaced by AI

We’ve recently been experimenting with AI generation and have been using Open AI, specifically Dall-E 2, so thought we would share our thoughts on what we’ve learnt.

For those of you unaware, Dall-E is an AI system that can generate photorealistic images and art from a description. To keep it simple, the best way to look at it is that Dall-E is an automated cook. You give it the ingredients and it will cook you up something. And this is why you’re not going to be replaced. Designers are chefs.

Something I always stress to people about design is that it's not linear. If someone says ‘Design me a logo’ - yes, you could just fart out the first thing that comes into your head; however there’s a few things that often get overlooked here. Firstly, the ability to visualise and actually come up with an idea is a skill in itself, and the best in their field have taken years to hone this. Secondly, how do you know that your design is going to be what the client is after off the bat? Being able to design a logo and design a good logo that is going to work well are two very different things.

My point is, to get the best results, not only do you need to know what ingredients to include, but also the order and treatment of them. As a designer (chef) we use a lot of tools (cooks) to turn our ideas into realities. Dall-E is the latest tool in our inventory, but it’s not going to do our job for us.

So how does Dall-E work — and is it just a gimmick?

Essentially you put in a structured description of what you want, including keywords to generate an image.

Go wild here.

Let’s put in a silly example to kick things off.

Typical day in the studio amirite? Weirdly, one of these AI generated people looks like someone in our studio 🤔

Let’s not dismiss how incredibly good this is. But you can easily see how this could be taken for a gimmick. Yes, it is fun to put in some ludicrous descriptions and see what it comes up with. But when you understand how to order those ingredients, you start to get some really interesting results.

I want to go.

As you can see from the above results, typography is one the weak points — along with detail — such as people's faces and features. However, should you need to work with the results, they are totally workable with other software such as Photoshop.

So what is it good for?

I guess the best thing to do is to talk about what we have found it useful for - and that is predominately conceptualisation and idea visualisation, as well as more abstract visuals.

Where it has its strength is in helping you to visualise something quickly from a conceptual point of view. Be that a photo, graphic style or some sort of composition. When you get it right it can be great for art direction, mood boarding or quick exploration to put something into a mockup.

A recent example was a request from a client where they needed a photo for a visual and hadn’t supplied any — as it was quite specific we were struggling to find stock image examples.

The fact these are AI generated and not real photos is pretty mind blowing.

These weren’t perfect but enough for us to put a concept together to share with the client.

The keen-eyed of you would notice the weird toenail — something easy to edit in Photoshop if this were to be used.

What this achieved for us was a very quick visual concept to give to the client for them to give us more feedback and direction, rather than sinking time into trawling through stock imagery (we all know how long this can take). The feedback from this was that these shoes aren’t actually flamenco shoes (shock, AI shoes aren’t real)! But it gave enough of a visual direction that we were able to commission a real photo to be used.

Once you start to understand how to feed in the information to generate the results, you can get some really good results, which not only could be used for conceptualisation but also as an actual visual.

Charlota from Web3co did a great deep dive on how she was able to curate some fantastic results, which you can read here.

Charlota's results

I can’t do detailed 3D modelling. DALL-E enabled me to find a workaround. You ofc don’t have precise control of details, tweaks and image quality. But I think this will inevitably come in the future.

The results are pretty impressive - but it takes a while to really know how to get the results you want. And even then, you really are taking a giant swing into the abyss.

So my job is safe?

For now… well, actually for a while. AI is and will always be a tool. Tools do not replace us, they improve our processes. Thus freeing up more time for more creativity and innovation.

When we use them well they can elevate the work that we are producing and give us the freedom to be more experimental.

I, for one, am excited to see how this technology develops, as I've seen some genuinely beautiful art created from it.

So if you’re worried that AI is going to steal your job, you need to think about who your clients are. A lot of platforms already exist out there for things like logo design generation, social media post creation, website building, etc - the list goes on. The people using those aren’t going to be coming to you in the first place. If they do, it’s our job as designers to communicate the importance of our value and the insight that we offer.