Dewan Mashuq Uz Zaman
Manchester City and Puma’s fan-generated AI kit shows how technology can open fashion design to more people. But as AI makes creation faster and easier, the bigger question is whether it is improving fashion or simply changing how it is produced and marketed.
Manchester City and Puma recently turned a fan-generated concept into a professional football kit. Puma opened its AI Creator to supporters, giving them an interactive tool to develop and refine kit ideas. Elliott entered with a clear concept inspired by Manchester City’s 2013/14 third kit and used around 10 prompts to develop it. He adjusted elements of the design and rejected versions that did not match what he had in mind. More than 180,000 designs were submitted before 10 finalists were selected. Elliott’s concept later won the public vote and became Manchester City’s 2026/27 third kit. Manchester City and Puma describe it as the first fan-generated AI kit to be worn by a professional football team.
The project shows that AI can change who gets involved in design. Whether it improves design itself is a different question.
From Fan to Designer
Traditionally, consumers enter the fashion process only after the product has already been designed. Generative AI allows consumers to take part earlier in the design process.
A supporter may know what a Manchester City shirt should look like without knowing how to sketch professionally or use advanced design software. AI lowers that barrier. Instead of merely telling brands what they like, consumers can develop and visualise their own ideas.
But participation is not the same as control. Puma and Manchester City still controlled the process. Fans could submit ideas, but the companies decided which ones moved forward and could still make changes before production.
Those submissions also gave Puma a large pool of creative input and could offer insight into the styles that attracted fans. The process therefore offered the brand something beyond participation.
The door may be more open, but the brand still owns the room.
Who Gets to Be the Designer?
The Manchester City experiment makes authorship more interesting than simply calling the shirt “AI-designed.” Elliott did not begin with a blank prompt. He already had a clear idea and used AI to visualise it. When the generated results did not match that direction, he rejected them and continued refining the design.
In that sense, AI acted more like a design tool than an independent designer. Elliott himself described the contribution as roughly “50-50” between his input and AI. Some logistical changes were made to the final design, including fading the middle section to accommodate the sponsor. Elliott said the technical details otherwise remained very close to his original concept.
This does not make AI irrelevant. It shows that the creative process was collaborative rather than automated. The harder question is whether AI is replacing design or simply changing the tools used to create it.
A quieter concern is what happens to entry-level design work. If AI takes on more of the early-stage tasks used to build experience, fashion may gain faster design tools while making it harder for young designers to develop their skills.
When More Becomes Less Distinctive
Generative AI can produce huge numbers of designs quickly, but quantity is not originality.
The Manchester City shirt has already drawn comparisons online with Chelsea’s 2003/04 away kit, which also featured a white base and a blue element running through the centre. That does not mean the newer shirt copied the older one. Football kits and fashion have always reused older ideas.
The concern is what happens when rearranging those ideas becomes almost effortless. If polished variations take seconds to produce, the difficult part is no longer creating options. It is deciding which one deserves to exist.
That could affect brands as well as individual products. Fashion labels and football clubs build visual identities over time. If many brands use similar generative systems drawing from large pools of existing imagery, they may gain speed while losing some of the quirks that make their work recognisable.
Mass participation creates another tension. Some memorable fashion designs were unusual when they first appeared. What seemed strange then later became part of their appeal. A process built around thousands of options and public voting could reward what the largest number of people already find appealing.
Retro kits show why that matters. Their appeal often comes from looking unmistakably like the period that produced them. A strange pattern or an unexpected design choice can become part of what makes a shirt memorable years later.
The safest design is not necessarily the one people remember.
The Costs Behind the Convenience
AI also introduces costs that are easy to ignore when the output appears in seconds.
One concern is environmental. Generative AI depends on data centres that require growing amounts of electricity and can also consume water for cooling. This creates an awkward contradiction for a fashion industry already facing questions about sustainability. The environmental cost of generating thousands of concepts that may never become products is therefore worth considering.
Another issue is authenticity. Research has found that consumers generally responded more favourably to human-designed clothing than AI-designed clothing, partly because the human designs were perceived as more authentic. In fashion, where the person behind a design can contribute to its value, that difference matters.
There is also the question of where generative AI gets its creative foundation. These systems can be trained on large amounts of existing copyrighted material, raising questions about whether creators have consented to their work being used and how they should be credited or compensated.
That creates an uncomfortable tension for an industry that places considerable value on protecting original designs.
Design Innovation or Marketing Innovation?
There is still a strong case for AI in the Manchester City project, but it may have less to do with the shirt itself.
The scale of the project turned supporters from spectators into part of the creative process.
That is an impressive marketing achievement. AI did not merely help create a design. It created an activity around the design.
If a conventional design team could have produced a shirt that supporters liked just as much, then AI may have transformed the campaign more than the product. There is nothing wrong with that. Marketing has always tried to bring consumers closer to brands.
The problem begins when technological novelty is treated as proof of creative progress.
Where AI Actually Fits
AI has a convincing place in fashion when it gives people an opportunity they previously did not have. A supporter without professional design skills could contribute an idea that eventually reached the pitch. That is meaningful.
But accessibility should not require pretending that the old process was broken. Human designers still bring judgment shaped by their experience and understanding of culture. They also remain accountable for the choices they make. AI can widen participation and speed up experimentation without becoming the reason a product exists.
The strongest argument for AI in fashion may therefore be relatively simple. It can help more people express ideas they were previously unable to visualise.
That is different from saying fashion needs AI to be creative.
What Will People Actually Remember?
Manchester City’s third kit will eventually face a test no algorithm can complete.
Years from now, supporters may remember what happened while the shirt was being worn rather than the prompts and generators behind it. If that happens, its AI origins may become little more than an interesting footnote. Like many older kits, its value will have come from the history attached to it.
If the most memorable thing about the shirt remains the fact that AI was involved, however, that tells us something very different.
Technology can make design faster and open the process to more people. But fashion was never valuable because it could produce the most options.
Fashion does not need more designs simply because machines can produce them. It needs designs worth remembering.
