I'm repeating a comment of mine from another thread, but this is not true. Both recording the audio of a conversation that you aren't party to and deriving biometrics from video without consent are both broad categories that are regulated depending on the state you're doing the "public usage" in.
If it was also illegal to (for example) input a photo of someone non-consenting into any kind of AI model or post it to any other online service? Then I don't see much problem.
People normalised installing spy doorbells, so every doorstep is centrally available to large organisations who want to do harm (government, amazon, meta, whoever)
I can't install a Ring doorbell if it takes a picture of the street outside my house. That was preexisting regulation (about surveillance cameras requiring permits for public spaces). Of course, people who now install Ring doorbells DO often record the street. But that's more a matter of enforcing the law.
It's easier to ban ring from selling devices in 2010 when nobody had them, then to take them away from millions who feel their personal benefits of not having to get off the couch to see who's at the door outweighs the societal harm.
That's before the arguments about societal benefits (coperganda does well at this). You change the argument from a hypothetical "this could help stop crime" to a concrete example "in this case we found out who robbed little old granny thanks to our surveillance network".