If AI can translate language-to-language with audio-to-audio in realtime.

Then couldn't RFK make a model of his 'pre-spasmotic dysphoria voice" (1996) using old recordings, and auto-translate his current voice to his old voice, making him more understandable while also showing people that he is up with AI tech?

(I would think training it on his current unique voice and it's tremors/inflections would tighten the translation too, helping to account for his uniqueness.)