A new approach to AI data puts communities in charge
microsoft ·
Microsoft has introduced the Community Library Creator, a research platform that helps communities shape how they are represented in AI-generated imagery. Instead of relying on data scraped from the internet, which can contain harmful stereotypes or gaps, advocacy groups use the tool to create their own curated collections of images and videos with descriptive annotations. The process involves community members identifying key themes, annotating up to about 400 real-world images, and rating AI-generated outputs to create a feedback loop for better representation. Crucially, the community owns the data through its advocacy organization and controls if and how it is shared, with consent and removal options. The initiative aims to give people more influence over AI development by putting their lived experiences and rights at the center, though it is currently run in a controlled way with specific organizations.
microsoft