AI has no soul, but it can extend ours

Lately a lot of people have been making AI images, especially in the "Studio Ghibli style", and it's quite controversial.

I'll skip the praise. The most common criticisms are "these pictures have no soul" and "they don't respect the original work".

I understand that, and partly agree. But I also want to share what I've learned from actually using AI at work over the past few months.

I lost to GMV MAX

In my company, I handle digital marketing myself. Six months ago, the major e-commerce platforms started rolling out "GMV MAX" ad systems, which at heart use deep learning to optimise ads.

I tried it at first, but gave up before long. It was too uncontrollable.

I spent money without knowing who it went to. Where were the ads showing? Results were summed up in a single ROAS number.

And I have to admit I had a bit of ego at the time:

"I've been running ads by hand for so many years. Am I really going to lose to an AI system?"

Reality slapped me quickly.

Over the next few months, ROAS kept sliding. We kept trying, tearing campaigns down and rebuilding them, switching models every week. Nothing worked.

Then in February I went to a platform event, and after a long conversation with our account manager I realised the platform was going all in on GMV MAX. That was when I really started to understand the logic behind this AI ad system.

I started trying GMV MAX again. At first I still wasn't happy, mainly because I wanted too much control:

"Running ads isn't spending money and hoping for luck. I want to know where every cent goes!"

Where is my edge, really?

But gradually I saw a more fundamental question:

"Where, actually, is my edge?"

The logic of manual ad buying goes like this:

I have a product. I need to find exactly who will buy it, then put the ad precisely in front of them.

But in front of AI, that logic no longer holds.

Because AI is far better than me at "finding people".

So what advantage do I still have?

We then made an important shift in our decisions:

Since AI wins at finding audiences, we'd go back to what we're really good at: the product itself.

We started putting more time and resources into refining our products. We also realised a good product isn't just about "value for money" (cheap and practical), but about being worth the price (something people are happy to pay for because of the quality and experience).

If you only compete on low prices, you fall into a vicious cycle.

A good product, plus AI precisely finding the right people, is the best combination.

Does AI have a soul?

Content is the same.

Making a Ghibli-style image with AI is nothing special now. Anyone can click a few buttons. But what truly moves people was never how polished the drawing is. It's whether it has feeling.

The Ghibli-style picture of my kids is, to me, a warm and precious piece. Technically it may not be perfect, but it has emotion, a story, and it resonates. To me, that is "quality content".

I've also come to understand something:

Quality content only has real value when it's in front of the right audience.

So, "does AI have a soul?"

I don't think AI itself has a soul.

But when we take products, content and brands we've poured our heart and soul into, and use AI as a tool to put them precisely in front of the people who truly care, AI becomes an extension of our soul.

AI isn't the lead. AI is the tool. People are the soul that truly gives it value.