There's a decent chance you've already been influenced by someone who doesn't exist.

Not a bot. Not a catfish. A fully AI-generated "person" — with a face, a lifestyle, and brand deals — created specifically to sell you something. And nobody told you they weren't real.

What happened

Brands across fashion, tech, and lifestyle are quietly using AI-generated influencers on social media to promote products. These aren't obviously synthetic characters with blue skin and anime eyes — they're designed to pass as real people, with curated feeds, relatable captions, and the kind of aspirational lifestyle shots you'd scroll past without a second thought.

Consumer groups are pushing back hard. The EU's Artificial Intelligence Act will require clear labelling of AI-generated content from August. But the UK currently has no specific rules mandating disclosure. Some content creators have reportedly been asked to sign NDAs to conceal their involvement in producing this synthetic content.

The telltale signs? Sometimes a hand with six fingers. Sometimes nothing at all.

This isn't just a marketing problem

Here's what caught my attention: this is a design story.

Someone designed those influencers. Someone chose the skin tone, the hairstyle, the wardrobe, the lighting. Someone decided what "authentic" should look like and then manufactured it pixel by pixel. That's design work — and it comes with design responsibility.

We've spent years talking about inclusive representation in design. Real bodies, real skin tones, real diversity. AI-generated influencers let brands sidestep all of that by creating "people" who are optimised for engagement rather than grounded in reality. They can be any ethnicity, any body type, any aesthetic — not because the brand values representation, but because the algorithm says it converts.

That should make us uncomfortable.

The broader trend is clear: AI is generating more and more of the visual layer between brands and audiences. From product mockups to social content to entire brand personas. And as designers, we're increasingly the ones building, directing, or at the very least enabling that layer.

The question isn't whether AI-generated content will become more common — it will. The question is whether we're going to insist on honesty when we help create it.

I think we should.

This connects to something else happening in design right now — the push to make our systems and processes more AI-ready. Not in a hype-driven way, but practically. If AI is going to generate more of our outputs, then the structures we build — design systems, component libraries, documentation — become the guardrails that determine whether those outputs are coherent or chaotic, honest or deceptive.

Tool spotlight: making your design system AI-ready

On a more practical note, there's a growing conversation about preparing design systems to work effectively with AI tools. The core idea is straightforward: AI tools that generate UI — whether that's prototypes, layouts, or component suggestions — produce dramatically better results when they have well-structured design systems to draw from.

What does "AI-ready" actually mean in practice? Three things:

  • Clear token naming. Your colour, spacing, and typography tokens need consistent, semantic names that an AI can interpret. colour-primary-500 tells an AI something useful. blue-3 doesn't.
  • Component documentation. Every component needs clear usage rules — when to use it, when not to, what variants exist. AI tools treat your documentation as instructions. Vague docs produce vague outputs.
  • Contextual rules. Define the relationships between components. Which patterns go together, which don't, and why. This is what prevents AI from generating layouts that are technically valid but contextually nonsensical.

If you're working with any AI-assisted design or prototyping tools, this is worth your time. The investment isn't glamorous, but it's the difference between AI that supports your system and AI that slowly erodes it through drift.

What to do with this

Two things to sit with this week.

First: if you're ever asked to design a synthetic person for marketing purposes, ask the hard question — will it be disclosed? If the answer is no, you're helping deceive real people. That's not a grey area. That's a choice.

Second: take an hour to audit your design system documentation. Not for humans — for machines. Look at your token names, your component descriptions, your usage guidelines. Ask yourself: if an AI tool consumed this tomorrow, would it produce something coherent? If not, that's your next project.

The AI layer in design is growing whether we're ready or not. The designers who'll do the best work aren't the ones who ignore it or blindly embrace it — they're the ones who build the structures and set the standards that keep it honest.