Researchers developed a “fashion map” using Facebook AI
Future. Researchers from Meta and Cornell University have developed an “underground fashion map” using Facebook’s AI, which can categorize the dress trends of 37 cities. While the tool could be a useful asset for visitors looking to get a grasp on new neighborhoods they plan on visiting, the tool could also be used by brands to laser-focus targeted ads… for better or for worse.
Researchers have cracked the dress code.
- The researchers created an “underground fashion map” across 37 cities.
- It was developed using “public, geolocated social media photos” from Instagram and Flickr.
- The map has the functionality to reveal “groupings” of people within those cities, labeling areas with monikers like “trendy” or “progressive.”
- Using an AI tool called GeoStyle, it can “forecast” fashion trends in a given location.
The report details that the fashion map can be used to help people determine where they want to visit, delineate between shopping and tourist areas, and figure out which areas are best for outdoor activities.
One of the co-authors of the study, Kavita Bala, sold her AI product-recognition company GrokStep to Facebook in 2020… so she has a history of working with Meta to leverage AI. And that’s an issue that some critics have with the AI tool. By working in conjunction with Meta, the tools can now be used by the company to “refine demographically and geographically targeted ads.”
Meta was sued by the ACLU when it was discovered that some finance, insurance, and real estate companies used current targeting tools to exclude some racial and gender groups. Monash University professor Jathan Sadowski explained the phenomenon as “weblining, algorithmic discrimination, or automated inequality.”