Why detecting deepfakes is the wrong problem to solve

And what the alternative architecture looks like I have been following deepfake detection research for the past year, mostly out of frustration. The frustration is not that the research is bad. Some o