Recall, if you will, the beginning of the classic 1995 film Clueless. Leading lady Cher Horowitz interrupts a montage of her and her fellow Beverly Hills peers zooming by in a Range Rover, laughing and flirting with a waiter through seductive milkshake consumption, and lounging by a decadent pool, with her characteristic Valley girl drawl: “Ok, so you’re probably going, ‘Is this, like a Noxema commercial, or what?!’” Cut to Cher’s room as she prepares for another day of pristine teenagedom. She turns to her once cutting-edge computer, activating a program entitled ‘Cher’s wardrobe’ that enables her to try outfits on her digital self. The program scans the ensembles Cher puts together, indicating whether or not her creations are a ‘match.’ The resulting outfit produced by the software is certainly that: the iconic yellow plaid jacket and skirt set that has become synonymous with the movie.

This technology is astounding in its ridiculousness—she needed a computer program to tell her that an identical jacket and skirt match? Nevertheless, there is undoubtedly an appeal to receiving an instantaneous ‘expert’ evaluation of your OOTD. The new AI system Fashion++ aims to capitalize on this desire, manufacturing a real-life Clueless closet. Utilizing visual recognition tools like the conditional generative adversarial network (cGAN), the system analyzes various elements of the user’s uploaded outfit (color, pattern, shape, etc.) and suggests minor alterations to improve its fashionability. It garners the criterium for ‘fashionable’ from a database with over 10,000 images of outfits—compiled from sites for self-proclaimed ‘fashion enthusiasts.’ In turn, it determines unfashionable outfits by mixing components from ensembles in the same database, combining the least similar garments.

Fashion++ prides itself on its practical approach to outfit improvement. It works with the clothing you have, providing tips on how to optimize an ensemble through stylistic tips like tucking in a shirt or switching to boots instead of sneakers. While its core values of providing cheap and low-effort alterations are commendable, the technology ultimately corrodes the joy and integrity of fashion, undermining personal style. By delegating the task of dressing up to an algorithm predicated upon homogeneity, we concede our individuality. Furthermore, the system contains serious blindspots—vintage and non-western clothing being two notable examples. But even if the system expanded its definition of ‘fashionable,’ its fundamental flaw—not accounting for the user’s unique preferences and personality—would persist.

Perhaps if we all possessed Cher Horowitz’s wardrobe, an AI service that seeks to match our garments would not be such a concession. But for those of us without an endless supply of plaid sets, sartorial novelty is found in the creativity of human imagination, not an algorithm.