The quality of the fakes is atrocious. Their audio clips don't even sound like deep fakes. They just sound like really terrible amateur impersonators.
It’s clear that state of the art deep fake technology can do better than what’s presented here.
What if the whole point of this exercise is to instill (false) confidence in the most people that they have the ability to distinguish a deep fake from the real thing?
For the record, I don’t believe this is the case and I fully expect that subsequent iterations of this exercise will be far more difficult.