It's already happening with profanity where the profane words are simply dropped out, but it could also be implemented with any other speech.
The video software will give the editor the option of accepting or rejecting any recommended changes, including "accept all changes" or "accept changes automatically".
The "AI" software could also change the mouth movements of the speaker.
It can be called "Autospeak", like Auto-TuneTM for music.
Also they will remove anything that smacks of ‘white privilege’. You have to have at least one “Person of Color” depicted in your video.