Google's Wardrobe Tool Fails to Deliver on Promise
A recent experiment with Google's new Wardrobe tool in Photos revealed some concerning issues. The AI-powered feature is designed to create a digital wardrobe based on clothing items found in users' photos, allowing them to try on virtual outfits and mix-and-match pieces.
The author, who tested the tool, was initially skeptical but saw some promise in its potential. However, the results were often hilariously bad and sometimes bizarre, such as an image of the author with a broken arm. The AI also struggled to find adequate clothing items from their photos, pulling only six pairs of pants and three hats from nearly 3,000 images.
The author's wife had similar issues, with the tool taking over three hours to build her wardrobe and ultimately finding only four items: two necklaces, one top, and one hat. Despite its limitations, the Wardrobe feature did produce some decent virtual outfits, but the novelty quickly wore off due to its inaccuracies.
Google notes that results may be unexpected, which is evident in this experiment. The author acknowledges potential in the concept but emphasizes that AI try-on tools still need significant improvement.