It's just as farfetched to think they're doing constant image/voice recognition with your phone's processor without you knowing. Those models are big and take a lot of computing power

It doesn’t need to be constant or continual. I don’t talk continually near my phone, for one.

Siri and other voice rec is able to get words from my audio without appearing to affect my phone function or slow it down. There are plenty of times my phone is idle that it could be doing that.

Same with photos. When you give many apps access to your photos, as we know, they will analyze and/or transit the metadata of all photos on the device. It’s not hard to imagine going farther than that, and as noted, processing or data transfer could be deferred.

Siri listens for a hotword, which is tremendously easier than running general speech detection models all the time.

Seems like you have to do at least minimal analysis of all sounds to be listening for a hot word.

Exactly, that's why "OK Google" isn't a thing.

There's an enormous difference in model complexity between key phrase detection and (general) automatic speech recognition + natural language processing. There's a reason nearly everything Google Assistant does except "OK Google" is done on Google's hardware.

> They don’t have to send back all of the raw data, just analyses like ‘person X’s face detected, is in a noisy place’ or ‘conversation included these 5 keywords’.

Not without making a noticeable drain on the user's battery.

When you fire up an app like Instagram, it does many things that create various drains on the battery. It opens the camera repeatedly when I don’t intend to use it, for instance. When you click stories, it opens the camera, often facing you. Who’s to say it doesn’t send that pic or some analysis of it to FB? I don’t think that would use any significantly higher battery or network anything especially since the processing and data transfer could be deferred.