What happens when you let an algorithm design a football kit? Apparently, you get something that looks like a compression garment had a fight with a heat map and nobody won.

Nike's new AI-assisted "computational design" kits for the 2026 World Cup have landed — and the internet is having a field day. The reaction isn't just mild disapproval. It's mockery. And honestly? It's deserved.

What happened

Nike unveiled national team kits for the 2026 World Cup that were created using what they're calling a computational design process. The approach uses performance data and body-mapping to determine how the kit fits and where materials are placed on the body. In theory, this should produce the ultimate performance garment — optimised for movement, ventilation, aerodynamics.

In practice, the kits look bizarre. Ill-fitting. Awkward on actual human bodies. The designs prioritise measurable performance metrics over the thing that football kits have always needed to do first: look good on people.

Players look uncomfortable. Fans are unimpressed. The algorithmically perfect kit turns out to be aesthetically terrible.

The bigger picture

This is a case study in what I'd call the optimisation trap — and it's something we should all be paying attention to as AI creeps deeper into design work.

Here's the core problem: AI is genuinely excellent at optimising for things you can measure. Airflow. Muscle compression zones. Material weight distribution. What it can't do — at least not yet — is account for the things you feel. Cultural meaning. Visual harmony. The way a kit looks when a player wheels away after scoring in front of 80,000 people.

Taste isn't a metric. You can't put it in a spreadsheet. And that's precisely why it's so hard to automate.

In design, we talk a lot about the balance between form and function. Nike's kits are a vivid reminder that when you hand function entirely over to a machine and strip out the human eye, you lose something essential. The craft of tailoring — of knowing that a collar should sit just so, that a stripe should hit at a certain point on the torso, that certain colour placements simply feel wrong — that's accumulated human knowledge. It's pattern recognition of a kind that current AI tools don't possess.

And this isn't just a fashion problem. The same tension exists in UI design, branding, illustration — anywhere that computational tools promise to optimise your work. An AI can generate a layout that meets every accessibility guideline and still feel lifeless. It can suggest a colour palette that passes every contrast check and still look ugly.

Optimisation and taste are different skills. We need both. But only one of them is currently human.

A counterpoint worth noting

While Nike was getting roasted, KFC quietly rolled out a rebrand that does the opposite. Subtle refinements to the Colonel Sanders logo. A redesigned bucket. Nothing revolutionary — just careful, considered evolution of a visual identity that's been working for decades.

No algorithm decided how to redraw Colonel Sanders' face. A designer did. And the result is exactly what good design evolution looks like: contemporary without being disruptive, fresh without being unfamiliar.

The contrast is striking. One brand threw computation at a design problem and got laughed at. Another trusted human judgement and got it right. That's not a coincidence.

Tool spotlight: your own design judgement

I know — not exactly a downloadable plugin. But hear me out.

As AI tools get better at generating, optimising, and suggesting, the most valuable skill you can develop is the ability to look at AI output and say "no, that's not right." Not based on data. Based on taste, experience, and an understanding of context that no model currently has.

This means actively practising visual criticism. When an AI tool generates a layout, don't just check whether it works. Ask whether it feels right. Train your eye the same way you'd train any other skill — by looking at a lot of work, forming opinions, and learning to articulate why something succeeds or fails beyond the measurable.

The designers who thrive alongside AI won't be the ones who accept its output uncritically. They'll be the ones who can edit it. Who can spot when something is technically correct but aesthetically dead.

Takeaway

Next time an AI tool hands you something that ticks every box on paper, pause before you ship it. Ask yourself: would I actually wear this? Would I put this on a wall? Does this make me feel anything?

If the answer's no, trust that instinct. It's the one thing the algorithm hasn't learned yet — and it might be the most important skill in your toolkit.

Nike just spent millions proving that optimisation without taste produces something nobody wants. Let's not make the same mistake at our desks.