Follow me

AI Can Create Influence. But Can It Create Belief?

AI Can Create Influence. But Can It Create Belief?

I have been in advertising since 2005. Two decades of learning, executing, getting things wrong, and slowly getting them right. Somewhere along that road, without any single moment marking it, I became an authority on the subject. I don’t say that to flatter myself. I say it because of what it cost, and the cost is the entire point of this article.

Here is what that authority actually looks like in practice. If I am sitting in a public place and I make an offhand comment about a campaign, a brand, or a piece of work, that comment carries weight. It gets taken as either criticism or endorsement, and a certain followership will accept it and act on it. Not because I demanded that trust, but because I earned the standing to be listened to over twenty years of visible, accountable work.

Now hold that image in your mind, because I want to use it as a measuring stick. I want to hold the AI influencer up against it and watch what happens.

The authority isn’t the opinion. It’s the twenty years of accountable work sitting behind the opinion and the reputation I would lose if I were casually wrong.

What an Influencer Is Actually For

Strip away the flattery the industry wraps around the word, and the purpose of an influencer is precise: to convert trust into action. A follower feels they know this person, not from a distance the way you know a film star, but intimately, because they’ve watched them in their kitchen, in their car, mid-thought. That felt closeness is the asset. And the influencer’s job is to point that trust at a product and have it convert, because a recommendation from someone you feel you know moves people in a way no polished advertisement ever will.

But, and this is the part that matters, that trust only means anything when it rests on real, lived experience the audience believes is genuine. The influencer used the product. Was disappointed by the alternative. Stood behind the thing and would answer for it if it failed. The recommendation is worth something precisely because there is a real person who bore the experience and who has something to lose if they misled you.

Where the AI Influencer Fails… Completely!

So ask the simplest possible question of a synthetic persona: how does an AI entity experience the brand? It doesn’t. It can’t. There is no self there to have an experience. It never used the product. It was never disappointed by a rival. It cannot be delighted, frustrated, surprised, or let down, because there is no one inside it to feel any of those things.

This is why an AI influencer is not simply a weaker influencer, a low-credibility version of the real thing. It fails the test by category, not by degree. Everything that gives an influencer their power lived experience, accountable judgment, a reputation staked on being right is exactly what a synthetic persona structurally cannot possess. What you are left with is a beautiful, controllable object that has been placed in front of you to sell. That is not a testimony. It is advertising copy in a human-shaped wrapper.

Simulated experience with no accountability behind it is just advertising copy in a human-shaped wrapper. It is an ad pretending to be a testimony.

The Missing Ingredient: Skin in the Game

Nassim Taleb has a sharper way to name what the AI persona lacks: skin in the game. My comment in that public place carries weight because everyone in the room knows what it would cost me to be wrong. I can be challenged, discredited, proven mistaken in front of the very students and professionals who trust me. That exposure, the fact that being wrong is expensive for me, is what makes my word worth anything at all.

Real influence is costly signalling. Credibility was never the opinion itself; it was the accumulated risk behind it. Every time I endorse something, I am spending a fraction of a reputation that took two decades to build, and people trust that endorsement precisely because it is costly for me to spend it carelessly.

The AI influencer is the purest expression of influence without skin in the game that has ever existed. It risks nothing. It has staked nothing. Being wrong costs it nothing, because it has no reputation of its own to lose, only the brand behind it does, one layer back and out of view. It is engineered to collect all the upside of credibility while being structurally incapable of paying any of the downside. And an audience that understands what influence actually costs will feel that hollowness, even when they can’t name it.

THE THREE THINGS INFLUENCE REQUIRES

  1. Lived experience: they actually used the thing.
  2. Accountable judgment: they can be challenged and proven wrong.
  3. A reputation at risk: being wrong is expensive for them.

A synthetic persona meets none of the three. That is not a gap. It is a disqualification.

So Are They Useless? No… But Call Them What They Are

I want to be fair here, because the honest case is more useful than a blanket dismissal. A synthetic persona, used well, is a genuinely powerful brand asset. Owned, controllable, tireless, and this is the underrated part: it eliminates an entire category of risk. It will never be caught in a scandal, never post something that torches your campaign at two in the morning, never arrive with a depleted-credibility ledger you forgot to audit, never drift off-message. You own it completely. As a creative character, that is real value.

But notice what that makes it. A mascot. And a mascot is honestly artificial; nobody believes the Michelin Man is a real person, and that honesty is exactly why he is safe. He carries brand personality without borrowing human trust. The danger with an AI “influencer” is that it is doing something more slippery: it is artificial pretending to be human. It is built to blur the line, not to declare it. And that pretence is the precise point where the safety problem enters, because now you are manufacturing intimate trust with no real self underneath it.

A synthetic persona is a brand asset, not a trust asset. Whether it’s safe collapses into one question: are you being honest that it isn’t real?

So my position, as a practitioner, is this. Disclosed and clearly synthetic, doing creative and campaign work, an AI persona is not just safe; it is often safer than a human influencer, for all the reasons above. Disguised as a real human with real experiences of your product, it is a liability with a fuse on it. Because the moment an audience believes a synthetic voice is a real person, you have built your brand’s trust on a lie that has a discovery date. And when it is discovered and it always is, the recall doesn’t land on the influencer, because there is no influencer to absorb it. It comes straight back to you.

The Word Itself Is the First Lie

Which brings me back to where I started, to the standing I spent twenty years earning, and to why I refuse to hand the same word to a rendered face. Calling these things “influencers” is itself the opening act of the deception, because the word smuggles in an implied credibility they cannot possess. Influence is a function of accountable, costly, lived judgment. A synthetic persona has none of the three. It is a sales asset, a magnificent, controllable, scalable sales asset, and that is a legitimate thing to be. But it is not an influencer.

You can use them. You can use them brilliantly, creatively, and to real commercial effect. What you cannot honestly do is call them what they are not. “Synthetic brand character” is honest. “Virtual persona” is honest. “AI influencer” is already halfway to the lie and in a business whose entire currency is trust, starting with a lie is a strange place to begin.

An influencer’s authority is borrowed against a reputation they can lose. A synthetic persona borrows against nothing, which is exactly why it can influence no one who understands what influence actually costs.