[zaccus]

I mean, if I give you a voice recorder, let you record secret stuff on it and give it back to me, do you really think it's reasonable to expect me to never listen to it under any circumstances? Really? It's not a technically complicated thing to think through is it?

[danShumway]

There's a way of phrasing the arrangement that makes it sound obvious, but that's not how it's ever been phrased to consumers. I don't think that's a conclusion ordinary people are equipped to reach on their own.

To be frank, we live in a world where consumers are apparently too stupid to understand what the "www" in a URL means, or that signing into a browser is different than signing into Google, but are also expected to understand that the magical witch-lady that caries out intelligent conversations during Superbowl commercials has to rely on manual classification by low-payed contractors in a low-security facility.

If we in the tech community are annoyed that ordinary people don't realize that technology isn't magic, it's kind of our own fault.

I personally think privacy is a continuum, and while I recognize that technically a company might hand-process some data, I have expectations about how often they'd do it, and what the context would be.

Isn't it reasonable to expect that some tasks in a company (putting passwords into a database, scanning emails to update a calendar, etc...) shouldn't be handled by humans? If I tell an ordinary person that 'Google' is reading their emails, on average I think they're going to substitute 'Google' with 'a computer', not 'Greg from Sales.' Voice assistants and AI are the same way -- people assume the process is completely automated.

[cameronbrown]

> Signing into a browser is different than signing into Google

Experienced this one from a family member recently. Wasn't possible to get through annoyingly.

[ziddoap]

If you buy a voice recorder to record your voice, you obviously know it is recording your voice.

My parents, bless their hearts, can't even grasp the difference between a browser and an operating system. You're telling me they should just know the intricacies of data processing through... osmosis?

Non-technical people may think that the data is being processed locally, if they even think about the data at all.

[CodiePetersen]

That's a nice strawman argument but it's not the scenario actually being played out. People may think that the talk to text option is no different than storing predictive text models on their own phone. That IS technically harder to think of because as I have stated most people don't know how that works and how much processing is needed for it.

There is a reason Facebook intentionally hid it's intent and usage and that's because participation would be markedly lower. Their motto is ask for forgiveness later and to hell with user privacy.