Researchers use fluid dynamics to spot artificial imposter voices

Then the deepfakes take their technique to generate a more perfect anatomical model?