We recently removed our kids’ weekday TV diet, and they were expectedly not pleased. But, by the time I looked up from assembling dinner, they were outside with the neighbors doing whatever kids do when nobody has handed them a plan. They came home hungry, ate without fuss, and climbed into bed easily.

I think a lot about what technology’s effects may or may not have on us as humans, and the topic of AI at work is similar to the kids’ screen time. Our rules around screen time at home weren’t failing because anyone lacked good intentions; the rules just start to break down because the easiest path is open as we’re spending our own time working out, preparing dinner, ending work.

Removing TV, however, made room for the unplanned, slightly unruly stuff nobody schedules: playing, inventing, getting a little muddy.

The slow drift

Something similar is happening to creative work, and it started well before AI.

Adam Mastroianni has made a wide-ranging case that society has been getting less weird for decades. The evidence is everywhere: teens taking fewer risks, culture leaning on sequels, and design that looks the same wherever you go. Brand logos have converged on the same clean wordmark, and so have cafés, apartment buildings and websites. His explanation is that life is safer and more valuable than it used to be, so we protect it more carefully. Of course that’s good news but the unfortunate cost is that weirdness now has to be chosen on purpose.

As a designer, I see it. Nobody ever trained us to make everything look alike!

AI didn’t cause this drift, but it accelerates it in two ways.

First, it makes the average the default. Ask for a logo, an image or a paragraph from Gen AI and you get something plausible, polished but interchangeable. It’s the weekday TV of creative work: easy, always on, and not harmful.

Second, it removes the beginner stage. Being bad at something is how people develop taste. You make an awkward thing, see why it’s awkward, and try again. When the first steps can be skipped, fewer people learn to see well enough to make something worth noticing.

Awe, the feeling of encountering something that couldn’t have come from anyone else, depends on a person having taken a risk, struggled, and been a little embarrassed first. That’s why I’d call weirdness a virtue. Not because weird is good in itself, since plenty of it is noise, but because it’s where originality comes from, and originality is where awe comes from.

What I do about it

I give my teams a lot of room at the start. Designers, researchers and others can pursue solutions without boundaries first, and I don’t mind if they use generative AI along the way. What I care about is the end: the final solution has to look and feel original, and that comes down to taste.

Taste isn’t something you can replace, and it isn’t something you’re born with. Juniors build it by experimenting and getting honest feedback. That’s what design critique is for. Hearing from peers and a senior what works, what doesn’t, and when something is plain or lacks originality is how people learn where the guardrails are. It’s also where being bad turns into being good, which is exactly the stage I worry we’re skipping.

My own process comes down to one rule: start by hand, finish by hand. I sketch with paper and a pencil first, to empty my head of fresh, random, original ideas before anything polished crowds them out. Then I model them into mid-fidelity, and tools can help there. But AI output is only ever a first draft. The last step is mine, putting my voice back in, because without it the work feels false right away.

We’re all still feeling out when to reach for assistive tools and when to sit down with a pencil. What I’ve learned, at work as at home, is that the answer isn’t a stricter rule. It’s deciding what comes first and who gets the last word.

Back outside

None of this means banning screens, or AI. I use both. But protecting weirdness takes more than good intentions, because the easy path is always open and always faster.

So I try to keep a few spaces where the safe, average path isn’t the default: where the first version is allowed to be bad, the odd idea gets a hearing, and honest critique still happens. Weirdness won’t happen on its own anymore. Somebody has to open the door and send everyone outside.