A Santa Cruz restaurant called The Salty Otter recently found itself at the center of a small, very online storm. The owner used an AI tool to generate a logo, diners noticed, and the reviews turned hostile, not about the food, about the mark. Within days she'd swapped it for a hand-made design. Design Twitter, or whatever we're calling it these days, split predictably: one camp said human designers charge too much and AI can do "just as good a job for simple things," the other said the AI logo "looks like AI, pay someone."
I want to argue both sides are having the wrong argument, and if we in the industry keep having it, we're going to lose the more important one.
The quality debate is a trap
It frames the whole question as: is the AI output good enough? That question has a shelf life measured in months, not years. The current generation of models is already good enough for a huge swath of "simple" marks, and it will only get better. If our defense of the profession rests on AI's outputs looking worse, we've tied our value proposition to a moving target that's moving against us. Every designer who spent this year quietly enjoying the "AI logos look bad" comfort is going to discover that comfort has an expiration date, probably sooner than they'd like.
The actual defense is somewhere else, and it's harder to say in a comment thread than "pay someone." It's this: a logo generated in isolation, with no knowledge of a business's actual customers, its competitive set, or the fifty small decisions that make a mark legible at sixteen pixels on a delivery-app icon and legible on a hand-painted sign out front, is not a comparable product to a mark made by someone who sat with the owner and asked why she named the place after an otter in the first place. Superside's recent brand research put a number on something we all feel intuitively: generic AI output is, by definition, the average of everything, and a brand mark's entire job is to not be average, to hold a specific place in a specific person's head. The tool doesn't remember the brand between generations. It has no idea what the business stands for. That isn't a quality gap that better training data eventually closes. It's a structural gap between a generation and a relationship.
What we're actually selling isn't the artifact
Here's the uncomfortable part for studios our size: if the argument is "hire a professional because the output is better," we've already conceded that the artifact is the product. It isn't. The Salty Otter's owner didn't need a logo. She needed a decision she could defend to herself and to her customers about who the restaurant is, and a partner who sat in the room while she made it. That's the thing AI can't yet package. It's also, not coincidentally, the thing most of us are worst at pricing and worst at explaining in a proposal, because "judgment" doesn't fit neatly into a line item next to "logo design, 3 concepts."
If this is the year AI makes "a nice-looking mark" a commodity, the studios that come out ahead won't be the ones arguing the commodity is still worth eight thousand dollars. They'll be the ones who stop selling the commodity and start pricing the judgment, the accountability, and the relationship explicitly, as things a client can see on the invoice and understand they're paying for. The Salty Otter didn't need the design industry to win an argument about AI on its behalf. It needed a studio to have made the case, months earlier, for what a studio actually does, before the owner ever opened a prompt box instead of an email.
