Twitter proposes flagging deepfakes, but would only remove content that threatens harm

Open in new window