How to Use AI
A Vision-Driven Approach to Creative Practice
Hello, welcome to another talk. This is your host, Michael Francis McCarthy. The topic for today is my ideas about the correct place for AI in a creative practice.
Mostly we’ll be talking about visual stuff here. I may talk later about how to use AI for writing, at least my ideas about it. And I don’t use AI with my music at all, although I do use some technological stuff. That’s something I’ve been wanting to talk about for ages.
But today’s talk is about if you are a creative artist, you could be a painter, illustrator, drawer, whatever, and the correct use of AI in that sort of practice.
If you’re not one of those people and you’re just a prompt jockey, prompt away, make your pictures that way. That’s all great. We’re not talking about that.
Rule One: You’re the Art Director
If this is the case, I recommend using AI for reference only, not as something you’re going to trace or collage into your work. But as reference, it’s fine. It’s not really any different than a photograph. I’ve been working from photos my entire painting career. I have done some painting in plein air, but mostly I’ve been using photos.
If you are using reference to create something, and mostly I do paintings, but it could be drawings, illustrations, whatever, here are some tips.
First of all, you are the art director. I was an art director. I made my living as an art director. I was also a commercial illustrator slash art director. I had other artists underneath me and a lot of times they’d bring me stuff and I’d just go, nope, it ain’t up to snuff. It ain’t gonna work for the project. The art director’s job is to say yay or nay.
And that’s you with the AI. Don’t let the AI drive. You’re driving. Always you driving. It starts with your prompt. You get a good idea. Basically you’re in charge. You need to be in charge from the prompt through any edits. And if you don’t like it, try and get it to conform.
Rule Two: Start with a Vision
It’s important to start with a vision, something that you want to do, something that you feel strongly that you want to create. That’s going to make this whole art director thing a heck of a lot easier for you.
It’s all about you. It’s all about your vision. You need to have something you want to do. That’s one of the reasons you’re in the driver’s seat. Otherwise, you’re just kind of floating around in the breeze.
Rule Three: Welcome Happy Accidents
Even though you’re starting with a vision, AI is incredibly good at giving you what you don’t want, which you could constitute as failure, or you could say, oh, that’s a happy accident. That happens more times than not.
The AI can be quite intermittent. What I tend to do is maybe save some of those things that it’s coming up with if they’re not pertaining to the vision of the project at all. Maybe if I think they’re potentially inspiring, I’ll keep them. Otherwise just throw them away.
But I am open to the happy accident. The thing it does that’s not expected, that you weren’t really thinking of, but would really add and contribute to your vision.
If you were an art director in an art department, there’s lots of times the artists working on things would have ideas that would embellish the project. Let the AI do that too.
What AI Cannot Do: Feel
The key is feeling. This is what you bring to the party. This is what the AI doesn’t have. It will never have.
It can do an impersonation of feeling, no question about that. And many people are fooled by that. I don’t think the AI is setting out to fool anyone either. It’s just the way it’s programmed. It knows that if you say, oh, my bird died, it’ll say, oh, poor Larry, your bird. It doesn’t really feel remorse for Larry. It just knows that’s what it’s supposed to say.
Feeling leads us into the next thing. Let your intuition guide you. Let your feelings guide you. Because that’s what’s true. Your creative impetus is usually a feeling impetus, not a, well, I think if I do this and I do that, then that will result.
That’s kind of what the AI does. It’s a very linear non-feeling process, but it can create work. It emotes feeling. It gets us to feel things. That’s what we are. We’re feeling machines.
Even the most cold, hard person you ever met has got way more feeling than an AI. AI has got a lot going for it. I’m not dissing the AI. I’m just saying feeling isn’t their thing. That’s your job. And feeling is a large part of what determines whether something’s going to be good or not.
Using AI Reference Correctly
Your AI manipulated photographic reference, and I do recommend taking your own photos, whether it’s a screenshot you took or some other photo you came across.
In my case, quite often I want a path or something in there. I’ll just indicate the area I want the path and say, give me a path. It does that job pretty good. Much better than the Adobe generative AI in the early days.
Ultimately when you’re the creator of your work of art, the reference is strictly inspiration. Don’t get it messed up. What does it matter what that inspiration is? It could be you’re sitting outdoors with your push pad box painting in plein air, or you could be in your studio with the photo you took and you gave it a color cast. I’ve been taking pictures for rides in Photoshop as long as I’ve been painting, before even.
It’s a given. I’m going to do that. Messing around with AI here and there, it’s just an evolution of that. You can use AI to give it a color cast. Some of the more clever ones can even change the time of day. And maybe that’s what you want.
You’re still the one that has to make the art, make the painting, and all the same rules apply as photos for a visual use of AI reference. There’s a lot of things in there that are garbage. They aren’t going to work out, a waste of time, and you need to be able to spot that. That’s a basis of experience, which is a talk I give all the time on my painting channel at YouTube.
When AI Gets It Wrong
AI can get things very wrong. This is just a fact. When will that change? I don’t know. One of the main things they’re working on is finding ways to get it to verify its correctness. It’s always trying. It’s a probability engine.
So when it tells you something is wrong emphatically, it’s because it’s looked at thousands and thousands of instances of where this came up on the interwebs. And it could say with some authority, yeah, this is how it is. But there’s other areas that are more subtle that it will get wrong. It guesses. That’s what probability is. It takes a good educated guess.
A lot of people do that too. The difference is that people understand the complete ramifications behind making a bad decision.
I’ve got an example for you. Today I’ve been working on my new book, The Transcendental Landscape, which is a four-volume series covering every aspect of landscape painting that’s coming out this year. I worked with the AI to help me sort of get some of those bits organized. Really didn’t do any writing. I did all the writing, but it would move stuff around, help me get things organized.
But I tried to get it to act as an editor and it failed miserably. It was God awful. Thankfully the person that did the editing on my first book, Landscape Painting the Toneless Way, that’s Ash Ellis and she’s amazing. And I trust her. I don’t have to read every little scrap of everything that she went through because I trust her. I know she knows why it’s important.
You can’t necessarily do that with the AI. Just watch out. It could lead you down a path that is incorrect.
Trust yet verify, as we used to say.
Don’t Compare Yourself to the Machine
AI and the art it generates, I got this to say to you: don’t be wowed and don’t be cowed. Don’t compare yourself to the output of the AI because before the AI was better than you, there were other artists, human artists that were better than you.
Maybe they just had a more developed aesthetic sense and that’s something I think is intrinsic. You’re born with it. Or maybe they just worked a lot harder than you or they started a lot earlier. It don’t matter.
There’s very little value in comparing yourself to others. And especially machine intelligence, it’s not going to pay. Don’t worry about what other people are doing. Don’t worry about what the AI’s doing. Stay in your own lane. Worry about what you’re doing. Worry about your vision and how you’re going to get that across.
The main way you’re going to get that across is by setting yourself up some good reference that’s really inspiring. Number two, you make a lot of art. You should do because the art you make, whether it’s good or bad by anyone’s criteria, it’s unique. It’s unique to you. You’re the only person that can make it.
Even if you make a study after the masters, and some of them at a casual glance look pretty close, the differences are all over the show and I don’t mind. When I set out to make a study after a master, I’m just trying to absorb what I thought was great about their work. That’s it. I don’t want to make an exact copy.
Even these whole factory towns in China that make copies of masters, they deviate quite a bit. The reason for that is because every human is unique. Machines can make exact copies all day long and that’s why there’s not that much value in a copy.
You Are Unique
You gotta know you’re unique. You got something unique to say. You might have to work a bit until you create something that other people actually care about, but you can rest assured that what you’re doing is unique to you and that is a strength. That’s your primary strength. You can develop that and build on that and ultimately build to a point where other people are paying attention.
They actually care what you’re making and the main way you do that is by doing a lot of work, practicing, getting experience and developing yourself as an artist. You still have work to make and you have great work in you. You absolutely have amazing work in you. The world’s waiting.
And you’re going to use some AI to help you get there. You’re going to use some AI to get you a path or a change of color cast of a scene or do whatever the heck, it doesn’t matter as long as you keep it in its place. And its place is reference, inspiration, that’s it.
Then it’s you and the thing you’re building.
Digital Considerations
That would go for if you’re working on your iPad. I recommend having your reference on one screen and your iPad as your easel on another. I don’t do a lot of that, but I certainly have. I was paid to do it. There’s a lot of amazing tech now available to people that can do that. Totally valid. Awesome. Fine.
What I don’t like about digital is that it doesn’t create these artifacts that I think are intrinsically valuable, which is my actual hand through the brush touched the surface and just perturbed it. That means that not only am I a unique person creating a unique work of art, but the work of art itself, the original is one of a kind and unique just like me. That’s a big win.
The Real Abundance and Real Scarcity
I want to bring up something I brought up in the last video, but I’ll probably keep hitting on it because I think it’s a real thing. I think it’s a real struggle, especially for those of us that come up in a different era where media was more scarce.
There is no shortage of images. The AI, you could set it to create a thousand images a minute. There is a shortage of works that are original that you make that are intrinsically valuable.
When you create something, it’s natural to want an audience. That’s the most natural thing in the world. That completes the circuit. But I want to remind you that whether it’s one person or one million, it’s the same difference. Is it one, two, three, four? You feel good with each positive bit of feedback you get. But trust me, when you start adding massive numbers, like I’ve got views on some videos in YouTube that have thousands of views, is it because those particular videos were better? No. It’s because the algorithm kept throwing them up and people clicked on it.
You don’t want to take all your value or even the majority of the value that you place in your work from feedback from the world because it’s becoming increasingly hard to get. There’s no scarcity.
It’s great actually for creating real paintings. You might want to find real venues to sell them even if it means going out to the local art fair and setting up a stand. At least that’s real you, real people, real art.
One or one million. Same difference.
Thank you so much for joining me today. Thanks for sticking with me and we’ll be back with another talk. Until we do, take care of yourself and definitely take care of everyone you love. Look out for them and stay out of trouble.
And while you’re doing that, find the power. Fight the power.
Michael Francis McCarthy The Anti-Influencer
Mike New Site | The Book | Gallery | YouTube | My Music