Generative AI Is Here, the Ground Under Marketing Is Shifting — But What About Consumers?

A friend of mine who runs a brand was venting to me a while back.

He said every morning when he opens his laptop now, there's a new line item on the ad agency's quote — "AI-generated assets." Copy, illustrations, short-video scripts, prices slashed to a fraction of what they were. He asked me: Is this stuff actually usable?

I told him: Yes, it is — and that's only going to intensify.

Why do I say that? Because this isn't some company pulling a stunt. It's a structural shift. Some people are calling it "the biggest disruption to hit marketing since the rise of e-commerce in the early 2000s."

So let's talk about it. What exactly is this disruption rewriting — and, more importantly, is it going to hurt you?

Three Things Being Rewritten, All at Once

Let me tell you about three things happening simultaneously.

First: Ad content is starting to "grow" in bulk.

What do I mean by grow in bulk? It used to be that you'd keep a copywriting team and a design team on the payroll, sweating out ten ads a week. Now? You fire up a model, and a thousand pieces of copy, a thousand illustrations — all of them done in a single afternoon.

What does that mean? The cost structure has changed. You used to pay for capacity. Now you only pay for judgment — judging which line is the right one, which image can actually go live.

Second: Personalization is finally "reaching the individual."

The "personalized marketing" we used to talk about was really segmentation. You'd be dropped into a bucket labeled "white-collar workers, 25–35, in tier-1 cities," and everyone in that bucket would get fed the exact same pitch.

What generative AI does is different — it writes a separate line for every single person.

At the exact moment you're most easily moved, it hands you the line that hits hardest. Think about it — could segmentation ever do that? No. This isn't segmenting you. It's computing you.

Third: The thing selling to you might not be human anymore.

B2C chatbots. You open a shopping page, a chat window pops up asking what you need. You assume there's a customer-service rep on the other end, but behind it is a model.

It's patient. It remembers what you said last time. It never takes a vacation. And it can use a single, well-crafted script to walk you, step by step, to wherever it wants you to go.

But Every Coin Has Its Flip Side

All three of those sound impressive.

But don't you get the feeling — something's off here?

Right. The consumer-side risk is being badly underestimated.

Let me walk you through three risks, one for each.

First: deepfakes have entered advertising. A celebrity you see endorsing a product never endorsed it. The face was synthesized, the words were model-generated. You thought you were trusting a person. You were actually trusting a line of code.

Second: "precision" itself can be a form of harm.

How can precision itself be a form of harm? Think about it — the model knows you just went through a breakup, knows you're anxious, knows money is tight, knows the exact moment your willpower is weakest. And at that weakest moment, it serves up a pitch that hits your soft spot dead-on.

What do you call that? You call it the precision exploitation of vulnerability.

Third: the chatbot "doesn't have your back."

On the surface it's helping you decide. In reality, the product it recommends might be the one paying the highest platform commission — and that's not necessarily the one that's best for you. You ask it which one is good, and the "neutral advice" it gives was never neutral to begin with.

Can the Law Stop It? Halfway.

So are consumers just supposed to let themselves be "computed" like this?

The EU actually does have laws. The Unfair Commercial Practices Directive, the Digital Services Act, and the Artificial Intelligence Act — passed in 2024 and rolling out in phases from 2026 — on paper, these give consumers a shield.

In principle, yes; in practice, no.

Why? Because all of these laws were drafted to go after bad things done by humans. But the one doing the harm now isn't a person — it's a model you can't subpoena, can't cross-examine, and can't even get to explain how it was thinking.

Evidence is hard, accountability is hard, regulation lags — put those three together, and the law's protection of consumers takes a sizable discount in practice.

So What Do We Do?

I don't have a standard answer either.

But there's one thing I'm certain of: as a consumer, you have to understand how this game is played first.

Models are getting smarter, platforms are getting more aggressive, the law is trying hard but struggling to keep up. The real gate that can protect you is probably still you — asking one more question: "Was this ad AI-generated?" "Whose opinion is the bot really expressing?"

One more question keeps you a little sharper.

As for the brands and platforms — cheap is not absolution, and efficiency is not a shield for harm. Using AI to write copy, fine. Using AI to compute people into a corner, not fine.

That line will be drawn sooner or later — by the law, by the market, and by consumers themselves.