If people need to be 'controlled', then why can't AT&T use voice recognition to do exactly what Twitter does, and if someone says something offensive in a phone

If people need to be 'controlled', then why can't AT&T use voice recognition to do exactly what Twitter does, and if someone says something offensive in a phone call or text like "There are two sexes" or "All lives matter" then the person's phone can just be disconnected right? This is what Twitter does. Permanent ban for even the slightest offence. No, the solution is to make all censorship illegal on large platforms, because otherwise the unsolvable question of "whose standards" it is that determine "offensive" arises. Right now Twitter employees consider themselves this high court and they will ban you based on their own political leanings, which only 50% of the country agree with.