To create a more accessible internet, context matters

Stanford researchers explore how to improve AI-generated descriptions of online images for blind and low-vision users.