There's a statistic worth sitting with: most people now say they can identify AI-generated content, and they trust it less when they do. That's not a niche observation from design forums. It's a shift in how credibility gets read, and it has direct implications for anyone making decisions about their brand's visual output.
For the last few years, the direction of travel in design tooling was toward smoother, faster, more polished output. AI made it possible to generate a competent logo, a competent layout, a competent illustration in seconds. The ceiling for "good enough" dropped to the floor. Everyone could produce something that looked professional.
When professional polish is available to everyone instantly, it stops signalling anything.
That's exactly the problem. When professional polish is available to everyone instantly, it stops signalling anything. Worse, a certain kind of over-polished, over-symmetrical output has started to read as synthetic, even when a human made it. The uncanny valley isn't limited to faces anymore. It's showing up in layout, in typography pairing, in the particular blandness of AI-assisted colour palettes.
The signal that's replacing polish
What's filling the credibility gap is imperfection that reads as evidence of human labour: the slightly rough edge, the hand-set element, the deliberately inconsistent detail. Not sloppiness, but considered imperfection, which is a much harder thing to produce than either flawless polish or genuine carelessness.
This is why zine aesthetics, scrapbook textures and visibly hand-made elements are turning up in brand work that has nothing to do with nostalgia. They're not a style choice. They're a trust signal, doing the same job that a handwritten note does inside a mass-produced package: this had a person behind it.
Where this matters and where it doesn't
This doesn't mean every brand should look homemade. A financial services brand and a craft goods brand don't need the same amount of visible human texture, and overcorrecting into deliberate scruffiness can look just as manufactured as the AI polish it's reacting against.
What it does mean is that "clean and generic" is no longer a safe default. If your brand identity could be swapped with three competitors' and nobody would notice, that used to be a mild inefficiency. Now it reads as a credibility problem, because clean-and-generic is exactly what AI tools produce by default.
The businesses handling this well aren't rejecting AI tools. They're using them for speed and iteration, then deliberately introducing the specific, human, slightly imperfect choices that a generic model wouldn't make on its own: an illustration style with real quirks, a typeface pairing that isn't the obvious suggestion, copy with actual opinions in it rather than balanced neutrality.
The practical test
Ask this about anything your brand puts out: if someone suspected it was AI-generated, would it survive scrutiny? Not whether it looks nice, but whether it has specific, ownable, human choices in it that a prompt wouldn't have produced unprompted.
If the honest answer is no, that's worth fixing before the volume of AI-generated brand content makes the whole category start to blur together, which is exactly the direction things are heading.
The Brand Foundation Workbook walks a small leadership team through defining the specific, ownable choices that make a brand unmistakably itself, before any of that gets handed to a designer or an AI tool. Six chapters, free to download.