You’ve noticed it, haven’t you? Every second website you land on these days has gone purple. Every AI tool, every design platform, every bloody SaaS startup looks like it was spat out of the same indigo-soaked template.
There’s a name for this phenomenon, turns out. They’re calling it the “AI Purple Problem,” and it’s symptomatic of something far more worrying than bad colour choices.
I’m still chucking about it actually having a name. OMG!
Personally, I reckon purple died the day Prince did. That colour belonged to one of the most brilliantly original artists who ever lived, and now it’s been commandeered by algorithms that wouldn’t know originality if it slapped them round the face. But I’m getting ahead of myself.
How Tailwind CSS Broke the Internet
About five years back, Tailwind CSS set its default button colour to indigo-500. Nobody thought much of it at the time. But that single decision spawned thousands of tutorials, open-source projects, and developer resources, all using that same default without questioning it. When AI models scraped the web for training data, they found this purple-heavy code everywhere and learned a dangerously simple pattern: modern design equals purple gradients.
Let me say that again.
LLM scrapers found a lot of purple and then regurgitated a lot of purple.
That’s exactly what AI does. It regurgitates what it finds. It doesn’t come up with anything brand new.
AI doesn’t actually prefer purple. It’s just pattern matching, creating what it statistically predicts will be acceptable and safe based on what it encountered most frequently during training.
The feedback loop this created is bloody difficult to break. AI learned purple, generated more purple sites, which became part of future training data, amplifying the bias even further, resulting in more (you’ve guessed already haven’t you?) purple!
The Bigger Problem: Creative Monoculture
Purple is just the visible symptom. The actual disease is homogenisation, and it’s spreading across every format AI touches. Images, videos, podcasts, copy – they all start looking the same because AI models are fundamentally designed to produce the average of their training data.
They minimise error by creating outputs that align closely to the statistical mean of everything they’ve seen.
When millions of people use the same tools with similar prompts, they all get variations of the same patterns back.
It’s like everyone consulting the same style guide simultaneously.
Remember when everyone jumped on the Studio Ghibli style with DALL-E 3? It went from novel to oversaturated to utterly dead in weeks because originality got flattened into uniformity.
Research shows that whilst AI helps people produce more ideas, those ideas are fundamentally similar because they’ve used similar tools in similar ways. Hardly anyone asking for anything original or daring to inject some originality.
Damn.
Mike Krieger, Instagram’s co-founder who’s now at Anthropic, is sceptical that AI-generated content platforms can succeed long-term precisely because content needs to feel varied, not like something you’ve “kind of seen before”.
Why Sameness Actually Kills Brands
Homogenisation is the enemy of memorability, and memorability is the foundation of brand building. When everything trends toward the mean, the distinctive edges get smoothed out and voices disappear.
Your site might be perfectly functional and tick every usability box, but if it looks (and sounds) identical to 10,000 others, it won’t stick in anyone’s mind. Users develop design fatigue. Nothing stands out. Conversions drop because you’ve become forgettable.
There’s a darker problem brewing too. AI models are learning from the patterns of patterns. They’re trained on content that’s already been refined and homogenised by years of best practices and A/B testing.
As AI generates more content that gets published online, that same generic output (and more of it regurgitated already) becomes part of future training data, amplifying the convergence. Some researchers warn that if we don’t rethink how we’re using AI now, human creativity will flatline by 2027.
I think my content is safe from being regurgitated because I use the word ‘bollocks’ a lot!
The Defence of Purple (Fair Enough)
Not everyone thinks this is a crisis. Some designers argue purple was chosen for perfectly valid reasons before AI ever got involved. It signals creativity and innovation, which fits AI’s brand identity perfectly. It evokes futurism and sophistication. It differentiates from traditional “tech blue” used by older companies. Purple blends blue’s reliability with red’s warmth, creating trust with approachability.
Major players like Shopify’s Polaris design system and Linear’s “Magic Blue” aesthetic reinforced purple as the colour signifying trust and action in interfaces before AI tools picked it up. So perhaps the colour itself isn’t the villain. Perhaps it’s just become a convenient scapegoat for a much larger issue about how we’re using AI tools without enough critical thought.
Breaking Out Requires Actual Effort
The solution isn’t abandoning AI tools. That ship sailed bloody ages ago, and honestly, there’s no point pretending otherwise. The solution is using them differently. Some designers are now creating prompts that specifically force AI to avoid purple unless requested, using brand-specific colour tokens (Doh! Like, of course!) and nature-inspired palettes instead.
Hybrid approaches seem most promising, where human creativity shapes intent whilst AI handles structural heavy lifting. This means taking AI-generated drafts and injecting personality, refining voice, and adding specific examples that only someone inside your operation would know.
Some platforms now allow training on your existing content to capture unique voice rather than generating from generic patterns.
Treat AI as a tool for efficiency whilst keeping distinctiveness firmly in human hands. Because when everybody starts doing the same thing, nothing’s unique anymore. And in a world drowning in content, being forgettable is the fastest way to irrelevance.
Finally, unless you can step onto a stage in front of tens of thousands of adoring fans and make magic with your guitar like that wonderful little chap from Philly did, put purple back in its box and try – at least try – to be a little bit unique. Go on. I dare ya!