If AI Can Manufacture Influencers, Is There Still a Living for Real People?

A while back, an old friend of mine who runs a beauty brand invited me out for tea.

He asked me: with a budget this small, should I grit my teeth and sign one top-tier celebrity, or find a hundred small influencers? Or should I just have AI build an influencer for me — one who doesn't eat, doesn't sleep, and never implodes in a scandal?

I said, that question of yours cuts right to the heart of it.

The influencer marketing industry is being rewritten by AI end to end. How you find people, how you make content, how you measure results — all of it has changed.

Let's start from the beginning.

AI Isn't the Newcomer

Many people think AI only broke into the marketing world after ChatGPT went viral in late 2022.

No.

When you open TikTok, why does one quick stop turn into an hour? Because its "For You" feed is constantly recording your micro-movements: how many seconds you watched this video, which one you dragged back to watch twice. Instagram rewards the depth of interaction, while YouTube and Snapchat rely on behavioral modeling to decide what to push to you.

Behind all of this is machine learning. Long before generative AI went mainstream, TikTok, Meta, and YouTube — plus Kolsquare, a platform specializing in influencer marketing — were already using machine learning to analyze behavior, deliver personalized recommendations, and optimize campaign performance.

What generative AI has done is slam the accelerator on a process that was already rolling.

A Few Dozen Creators, and Human Hands Can't Keep Up

To understand the work AI is doing now, you first have to look at how the industry itself has changed.

In the early years, influencer marketing meant throwing big money at celebrities. And now? Brands increasingly prefer micro and nano influencers. What is a nano influencer? Maybe only a few thousand followers, but the comment section reads like friends chatting. Small following, deep trust.

Authenticity, at scale.

But the trouble followed. It used to be that one campaign managed one celebrity; now one campaign has to manage a few dozen creators at the same time: signing contracts, reviewing content, watching the data, writing reports. Do all of it by hand, and you'll work the team into the ground.

So AI has taken a seat in the command center. It can sift the right creators out of tens of thousands by audience quality, engagement performance, and brand fit; it can predict a campaign's ROI; it can pair different messaging with different audiences; and it can take over every repetitive chore. That is the core of what platforms like Kolsquare do.

And the time it saves? Give it back to creativity — back to the story itself.

Eight Out of Ten Creators Are Already Using It

And the creators themselves? Do they use it?

Kolsquare gave a set of numbers in its "Creator Economy" survey, and I'll read them to you: among European creators, 80% are already using AI tools, and more than half open them every week.

Using them for what? 72% to find ideas, write scripts, and edit videos; 40% for search optimization; 38% to dig into data; 34% to generate images and videos outright.

Honestly, that ratio is higher than I expected.

The tools are all off the shelf: Runway, Descript, and OpusClip do the manual labor of editing and clipping fast and cheap. This crowd has already turned AI into a creative partner; surprisingly few of them are using it to slack off.

But hold off on celebrating. The stronger the tools become, the harder one question is to dodge: where does the "human touch" that makes fans want to keep watching come from?

A machine can edit a video ten times faster.

But when a fan leaves you a message in the middle of the night, is it because they want a chat with CapCut?

Virtual Influencers: They Won't Implode, But They Can Wreck Your Brand

Now let's talk about the option that tempted my friend.

What is a virtual influencer? It is a "person" generated with AI: it has a face, a personality, an account — just no flesh-and-blood body.

This is not science fiction. Virtual humans like Lil Miquela and Shudu have already worked with brands like Prada and BMW. The industry had already predicted that by 2025, this market could reach $9.6 billion.

What's in it for brands? Control. Real people get tired and say the wrong thing; virtual humans don't. Want a few avatars? Have a few. Want a particular persona? Generate it.

But everything has its flip side.

Research has found a phenomenon that stings: when a virtual influencer flops, users blame the brand, not the robot. The logic is not complicated: the robot has no credibility — all of it is staked on your brand. Use its face, and you're the one on the hook.

So transparency is the bottom line. If the content is synthetic, the audience must be told.

There is also a finer distinction here. In spec-driven categories like tech and sports, AI creators' recommendations can genuinely beat real humans; but in emotion-driven categories like fashion and beauty, real humans still hold the upper hand. People trust the person first — and only then the words.

Platforms Have Started "Creating" Content Themselves

In 2025, two events are worth a line in the record.

OpenAI brought out Sora: one line of text generates a 10-second video; it can mix and cut the footage and the sound, and with the "Cameo" feature, it can put your own face into the scene. Meta's Vibes is more direct, packing these capabilities straight into Instagram and Facebook.

Notice that these two are different from the platforms of the past. The old platforms hosted content posted by users; these generate content themselves.

From here on, a new crop of AI-native creators will emerge.

You hardly need to guess whether these tools will take off.

The real question is: when even "the connection between people" is machine-generated, can users still find meaning in it?

I don't know.

But everyone who builds brands will have to answer it sooner or later.

Regulators Have Stepped Onto the Field

And don't think this is merely a debate about values — the hand of regulation has already reached in.

The EU's Artificial Intelligence Act is the world's first systematic AI framework, requiring transparency, human oversight, and risk-based classification. The UK has taken the principles route, emphasizing fairness and accountability. The US is advancing at the state level, with California and other states already introducing transparency bills.

In plain English, it comes down to two lines: if you use AI, say so; at the critical steps, a human must stay in the loop.

Back to My Friend's Question

Having talked this far, I can now answer my friend.

My judgment: AI will take over efficiency, but it cannot take over the human heart. It can help you find the right batch of creators, edit videos ten times faster, and run the math on a campaign down to the last decimal.

But that tap of "follow" — it's given to the living person on the other side of the screen.

AI is like a microphone. A microphone can't sing, but it can make the person who can sing heard by ten thousand people.

AI is the enabler. People are the influence itself.

As he was leaving that day, he said, I get it — the picking and the cutting go to AI first.

I said, right.

And then, keep yourself in front of the camera.