Deepfake Detection Dataset Aims to Keep Up With Generative AI

# Deepfake Detection Gets a New Tool Researchers from Microsoft, Northwestern University, and a nonprofit called Witness have created a large library of AI-generated fake videos, images, and audio clips to help computers get better at spotting deepfakes before they spread online. As AI tools become easier for anyone to use—even through smartphone apps—the ability to detect these convincing fakes is becoming increasingly important for protecting people from scams, fraud, and abuse. This new dataset is designed to train detection systems to keep pace with rapidly improving AI technology.