Harm Amplification in Text-to-Image Models

Open in new window