AI took over my closet!

Trying on clothes at the shop

Words and photos by Daniella Dogbe and Alex Gageanu

Personal style is individual to each person, as what we wear is intrinsic to our personalities, professions and our hobbies. Sometimes the decision can be tricky, however, and that’s where technology steps in.

With AI-powered styling tools promising to curate the perfect outfit, we decided to put them to the test. Can an algorithm really dress us better than we can dress ourselves, or is fashion still best left to us humans?

The Experiment: can AI style us?

We started by exploring AI-driven styling platforms, which claim to analyse body type, colour preferences, and fashion history to suggest the best outfits. The idea sounds futuristic, like having a personal stylist at your fingertips, but we wanted to see if the results matched up with real-world fashion.

To add another layer to our experiment, we analysed two different personal styles and imagined how AI might approach their wardrobe choices.

Here’s what Danjela, 20 had to say about experimenting with her fashion through AI.

We asked ChatGPT to style Danjela after taking a photo of her outfit and this is what AI suggested.

Style Profile 1: edgy & monochromatic chic

The first person embodies a sleek, trendy aesthetic with a preference for neutral colours, comfortable silhouettes, and bold makeup. Perhaps, AI might categorize this as “minimalist streetwear.” But could it capture the subtleties of of personal expression?

A suggested AI-generated outfit for this look might include:

  • Top: Oversized graphic tee or mesh long-sleeve top layered under a coat
  • Bottom: Cargo pants or oversized joggers
  • Shoes: Sneakers (Nike Dunks, Converse) or Combat boots
  • Accessories: Crossbody bag, layered rings and statement sunglasses

Next we met student Allison McKinnon, 29, who had a more sentimental approach to fashion. We asked her what she thought.

Then, we asked Chat GPT to style Allison after taking a photo of her outfit and this is what AI suggested.

Style profile 2: soft & vintage-inspired

The second person embraces a softer, more whimsical approach to fashion, think florals, pastel tones, and effortless layering, paired with unexpected elements like sneakers and denim. AI might recognize this as “cottage core” or “fairy core” but would it be able to balance the aesthetic correctly?

A complementary AI-generated outfit might include:

  • A puff-sleeve blouse or lace-trimmed top
  • A flowy midi skirt or corduroy trousers
  • Chunky sneakers or Mary Janes for contrast
  • Vintage-inspired accessories like layered necklaces and hair clips

The key question: Would AI lean too heavily into one category (hyper-feminine vs streetwear) and miss the broad mix that makes the look unique?

Can AI really replace human stylists?

While AI can create aesthetically pleasing suggestions, it still struggles with the nuances of personal taste, cultural context, and mood-based fashion choices. An algorithm might recognize a pattern in someone’s choice of clothing, but it can’t fully grasp why they gravitate toward certain textures, colours, or silhouettes.

For example, Mckinnon explained to us that the majority of her fashion she wears comes from her late grandmother. That is an essential human quality that AI doesn’t have: emotion.

Fashion is also about storytelling, which is something AI hasn’t quite mastered yet. While AI can be a helpful tool for inspiration, it likely won’t replace the human touch any time soon.

The verdict: AI as a fashion assistant, not a fashion guru

AI styling is fun to experiment with, and it can help decision-making for those who struggle with outfit choices. However, it’s not quite advanced enough to replace personal styling completely.

The best approach? Use AI as a helping hand while still trusting your own fashion instincts. After all, style isn’t just about what looks good, it’s about what feels right.

Would you let AI take over your closet? Let us know in the comments!

Be the first to comment

Leave a Reply

Your email address will not be published.


*