Fraunhofer researchers are developing a real-time warning system for deepfake video meetings, targeting scams where AI-generated bosses and coworkers pressure employees into costly decisions.