That's how deepfake AI has been trained so far - but as you can see, it's still flawed enough to notice.
This is another reason I'm not concerned. Deepfake AI may get better and better over time but I don't see a reason to think it will ever outpace deepfake-detection AI, which would get just as better over just as much time.
It is not about if it will be detected. It is about how many harm is there done before the deepfake is detected and stopped spreading (if possible at all). News and fake news travel with the speed of light (fiberglass) around the globe. And can be trending within a blink of the eye.