AI Is NOT a Magic Wand

The other day, my client said to me:

“I’m thinking of moving my website to Lovable. What are your thoughts?”

For those who haven’t heard of Lovable, it’s an AI-powered tool that lets you build websites and web applications by describing what you want in plain language. And if you don’t know me personally, you may know me as a performer. But when I’m not on stage, I run a web development business and take care of my clients’ websites.

So, yes, I have thoughts. Honestly, I have mixed feelings about AI. I’m not against my clients using AI. When you have a limited budget and limited resources, sometimes AI is the best solution.

But AI is not your magic wand.

You say, “Abracadabra,” and AI generates what you ask for. The problem is, what it generates isn’t necessarily good.

Oftentimes, I have to tell AI:
“This solution is inefficient.”
“This is wrong.”
“You completely misunderstood what I was asking.”
And sometimes I want to ask:
“Are you making these silly mistakes intentionally so I have to spend more tokens?”

Probably not. But I get suspicious. The point is: you cannot blindly trust AI.

I often see websites generated through Lovable that have flaws in SEO. And I’m not blaming Lovable. It’s not Lovable’s fault. It’s the user’s. People sometimes think that when they use AI, SEO will just work perfectly.

NO.

Lovable provides the tools. But you still need to review what it generates and make adjustments. You need to understand SEO. You need to test the website.

AI can build a website. That doesn’t mean it knows what makes a “good” website.

And then there’s the creative side.

We’ve seen incredible video clips and images generated by AI. At the same time, when there’s no humanity behind the process, the result can be cringingly obvious. Just because the group is related to performing doesn’t mean you should slap a microphone into the logo.

But that’s what happens when you think: “Performing group = microphone.”

AI doesn’t think: Who are these people? What do they believe? What makes them different? What story do they want to tell?

It generates something based on your request—and on what already exists. That’s why so many AI-generated things can look polished but still feel strangely familiar.

My creative partner-in-crime, Anthony Phills, creates amazing logos. And here’s what makes his work different. If you ask him about any of his logos, he can tell you a story about it.

The font has a reason.
The line has a reason.
The color has a reason.
The shape has a reason.
Every little detail has a story.

He doesn’t use AI to create his logos.

And that’s exactly why I think his work is such a good example of what AI can’t replace.

Anthony’s creativity doesn’t come from a prompt. It comes from his brain. From his experience. From his taste. From all the little observations and decisions that happen before he ever puts something on the screen. You can give AI a detailed description of what you want. You can explain the colors, the shapes, the feeling, the story, and the audience. AI can generate something based on those instructions and what already exists. But that’s not the same as having a reason for every creative decision.

AI can generate an image. Anthony creates a story that happens to become an image. And that’s a big difference.

The other day, I went to a Japanese restaurant. From the moment I walked in, everything looked great. The ambience. The furniture. The counter. The lighting. The plates. The glasses. Everything looked carefully designed.

Then I saw the chopsticks. They were nice chopsticks. But there was a paper band wrapped around them, and it was surprisingly difficult to take off. My friend suggested twisting the chopsticks to rip the paper.

Yes. It worked. But I thought: That’s a bad user experience. And then I started noticing other things. Every little detail was slightly off. Maybe they were Japanese food enthusiasts. Maybe they had done their research. But maybe they simply didn’t have a Japanese person to critique the experience. Then I looked around. There wasn’t a single Japanese person working there—which is nothing wrong with. But I couldn’t help wondering: How deeply do they understand Japanese cuisine? If you don’t know much about it, let me tell you something. Japanese dining isn’t just about the food.

The placement of the chopsticks matters.
The choice of plates matters.
The shape of the food matters.
The way the food is arranged matters.
The choice of a cup for miso soup matters.
The paper band for chopsticks matters.

Each element is subtle. But each element can have a reason. And when all those little things are right, you might not even consciously notice them. They work together like an orchestra, each instrument playing its part in harmony. But when they’re slightly wrong? You feel that, too. The hand rolls were shaped a little differently. The texture of the nori (seaweed) didn’t feel right. The choice of the miso soup cup felt wrong. The way they placed the pork katsu was a little off. Nothing was terribly wrong. And don’t get me wrong—the food was good. But everything was just…slightly off.

If AI created a restaurant, I imagine it would be like this.

It would have all the right ingredients.
It would look right.
It would taste fine.
But something would be missing.

I don’t think the future is humans versus AI. That’s not the interesting question. The interesting question is: How does AI help your audience’s experience? 

And if you don’t pay attention, AI can make you incredibly efficient at doing the wrong thing. That’s the scary part. AI may give you everything you asked for, but you still need to experience it as a user and ask:

“Something is off. What is it?”

That’s where humans still matter. Especially the picky ones…Like me.

While I’m accommodating AI into my life, sometimes I wonder: How much longer will I be able to review AI’s work? What happens if I keep relying on AI to write, design, code, and solve problems for me? What if, eventually, my own brain isn’t good enough to tell what works and what doesn’t?

Maybe that’s the scariest part. Not that AI will replace humans. But that humans might slowly lose the ability to tell when AI is wrong. 

What if one day, I walk into an AI-generated Japanese restaurant and don’t notice anything is wrong?

That might be the scariest thing of all.