The Right Tool for the Job

I have a relative who picked up a set of psychological tools through a self-help group. The problem is that they habitually reach for exactly the wrong tool for whatever social situation they’re facing and then wonder why nothing ever gets resolved. You’ve probably heard the expression, “To a hammer, everything becomes a nail.” This is adjacent to that, except in this case, the screwdriver is a hammer, the nail is a screw, and eventually, the whole construction project winds up on the floor, begging for mercy.

This kind of thing has been happening in the creative world for a long time.

 

The Logo Problem

Before AI, I regularly came across self-taught designers who saw no real reason to build a logo in Illustrator or any other vector-based tool. Photoshop could handle vectors, too, they reasoned. And besides, a logo built in Photoshop could be cinematic and dramatic.

For the uninitiated: a vector logo is art built from mathematical paths rather than pixels. That means it scales infinitely, up or down, without the edges going soft or blurry. Quality is always maintained. A Photoshop file can also do that if you place a vector object in the file, but now you’re sacrificing file efficiency. There is an excellent best practice of using vectors turned into smart objects in Photoshop that can be immensely helpful. But that is another article altogether. In short, Adobe Illustrator forces you to think in simple shapes, which helps in logo reproduction even in the digital space.

As long as the logo lives only in the cinematic world, you might get away with it. The trouble is, it rarely stays there. At some point, you’ll need to print: a T-shirt, a hat, a brochure, and business cards. And that’s when the problems start. Being able to reduce a logo to its basic elements is key.

Someone, usually not the original designer, now has to reverse-engineer the logo and rebuild it properly. Gradients don’t work with embroidery. Blurred edges, textures, and complex color blends often need to be completely reimagined depending on what the logo needs to do. And all of this happens while trying to honor the original design intent and manage client expectations that were set when they first saw it.

None of that would have been necessary if it had been built as a vector from the start. Do that first, and if you want a cinematic Photoshop version for digital use, go for it. Just be clear with your client about where it can and can’t go. Use the right tool for the job, and most of this headache disappears.

 

The AI Tool

It would be dishonest not to draw some parallels here.

Dozens of platforms will generate a logo, an ad, or even a full campaign from a well-crafted prompt. And yes, technically, you can produce several logo options in two minutes. I’ve run that experiment more times than I can count, trying to reach an honest conclusion I thought I might have to face.

I considered leaving the business entirely. If AI could genuinely do my job, why hang around watching it happen? As much as I love advertising and design, maybe it would be time to walk off the field and learn to love something else.

I kept thinking about Ol’ John Henry. He’s the man who won a railroad spike drilling contest against a steam-powered machine through sheer human will and strength. He beat the machine. But then he died of exhaustion. And then they used the machine later anyway.

So why fight it? Go be a firefighter, an electrician, a plumber. AI can’t do those things.

But then a persistent thought started to surface. Maybe it isn’t about beating the machines. Maybe it’s about beating your competition with the help of the machines. Maybe it’s about having an edge.

 

What AI Can’t Do

As I worked my way through the vexed state of that question, I was experimenting with logos, images, video, writing, campaigns, concept development, and more, one thing consistently held: just because you can do something a certain way doesn’t mean you should. It’s the same lesson as the Photoshop logo (and Jurassic Park).

Experience matters here. If you’re using AI for creative work and nobody on your team has deep creative experience, hire someone now. The more experience, the better. Otherwise, you’re just cranking out content that may not be tied to your brand look and feel or your brand voice.

Here’s why. Experience is about more than craft, although craft is a powerful differentiator, especially right now. Experience means decades of learning what works and what doesn’t. What makes people laugh? What pulls people into a story? What makes them stop scrolling?

Experience is about caring. Caring about what is on the page and caring about what the reaction will be. It isn’t an exact science. Sure, that unnerves people. But if you take it and try to make it into a science, suddenly the art doesn’t seem all that appealing anymore. Does that sound familiar? Of course, the retort would be “But I’ve seen a ton of artistic creations with AI that I loved.” Yes. But you are looking at something that originated in the mind of a human who had human experience. Not the AI itself, spontaneously creating art on its own. And don’t forget, AI was trained on other people’s art, which is mind-blowing in its own way. I’m not sure why anyone would think its ok to steal someone else’s IP. That is really unfair for the people who spent decades creating a branded look that belonged solely to them. To now be able to just type “in the style of” robs the people who owned that style and robs you of the joy of discovering your own style. The old masters in the art field used to copy each other’s work relentlessly. But this wasn’t in search of a perfect copy. It was to discover their own style through the reproduction of others. In other words, in that case, reproduction led to invention. And maybe that can be the case for AI, too. However, I see a disturbing number of people proudly displaying their AI work that directly rips off a style and are beaming about how close it is to the original. As one of my bosses used to say, “I don’t think you’re done baking that.”

So, to reiterate. Experience cares about quality. It cares about the brand’s image. It cares about longevity. It cares whether you succeed.

AI doesn’t.

I’m not being an AIsshole about it. It genuinely doesn’t care. At its core, everything is a one or a zero. Binary. Ask it! It will tell you that humans are responsible for setting trends, maintaining cultural relevance, refining aesthetics, and protecting brands. AI is just not interested in any of these things.

 

We Care, and AI Doesn’t

AI is a powerful tool. It can help all of us work faster and get more done. But it shouldn’t be asked to do more than that, because it doesn’t have the capacity to care. And it never will. That requires something AI will never have: real-world experience and humanity.

The Bottom Line

So the bottom line is this. I’m knee-deep in AI every single day, figuring out how AI can best serve the work. But I would never ask it to care about what I’m doing.

Because it won’t.

Randall Hooker
Executive Creative Director | Founder
Atomico Creative