Google Photos AI Wardrobe Fails Hilariously

Google Photos has introduced a new AI-powered Wardrobe feature that aims to create a digital closet from users' existing photo libraries. The tool automatically identifies clothing items to allow users to mix and match outfits and see how they might look. However, initial testing revealed significant flaws, including an instance where the AI bizarrely generated an image with a broken arm.

Despite having a decade's worth of photos, the feature struggled to build a comprehensive wardrobe. It identified only a limited number of items, such as six pairs of pants and two t-shirts, for one user. The AI also had issues accurately replicating some items, with logos and text appearing garbled.

When attempting to build a wardrobe for another user, the process took over three hours, ultimately yielding only four clothing items: two necklaces, one top, and one hat. This limited output, even after extensive processing time, underscores the current immaturity of AI try-on technology, suggesting it still requires substantial refinement.