Good point, but at least for now, video deepfakes can still be detected via machine learning. Also, deepfakes work well for replacing faces, but not yet for sy

Good point, but at least for now, video deepfakes can still be detected via machine learning. Also, deepfakes work well for replacing faces, but not yet for synthesizing actual videos of complex scenes (like crowds) : https://www.youtube.com/watch?v=IMZkLVBhcig https://www.youtube.com/watch?v=IMZkLVBhcig But then you're right, it won't take long before that's possible. Whether we can continue sniffing out deepfakes with ML is unclear.