Comment by ComplexSystems
1 year ago
I don't think any such conspiracy or secret-keeping is required; one need not attribute any of this to malice which can be attributed to incompetence. There already exists a system, on everyone's phones, which is listening to audio at all times, occasionally activates in response to a wake word and runs search queries and such from it. The point of this system, when used properly, is intended to take a recording of your voice, transcribe it into text, send it into a search engine, associate the query with your account and search history and use it to influence ad preferences in response to future queries. The functioning of this system involves sending data through an opaque and complex chain of custody, sometimes involving third parties, which - even if intended to comply with privacy and security protocols - could easily be mishandled either maliciously or accidentally, as happens in software development all the time. This includes but is not limited to:
1. The occasional false positive response to a wake word that wasn't really a wake word, causing search queries to be run that you didn't intend.
2. This data being accidentally mishandled behind the scenes in Apple's servers in some way, such as developer error leading to the data accidentally landing in the wrong folder, being labeled with an incorrect flag in some database somewhere, or otherwise being given the wrong level of sensitivity.
3. This data being deliberately "correctly-handled" behind the scenes in Apple's servers in some way that users wouldn't like, but technically agreed to when they first used the phone.
4. This data being used for valid "QA" purposes that, for all intents and purposes, include situations that user would probably not be comfortable with, but also technically agreed to.
5. An unforeseen security vulnerability affecting any part of this process.
6. Malware on the phone interfering with any of the above.
7. Not-quite-malware that you agreed to install, doing things you're not quite happy with but technically agreed to, which is somehow in the loop of any part of this process.
Again, we can debate all day which of these are true - hopefully none of them are true. But we're talking about software development here, where these sorts of things happen on a daily basis. None of this is "they faked the moon landing" kind of stuff, and all lead to the same result from the standpoint of user experience.
> The occasional false positive response to a wake word that wasn't really a wake word, causing search queries to be run that you didn't intend.
Nobody ever describes that behaviour, though. They don't say "a) we had a conversation, b) our smart speaker suddenly interrupted and gave us some search results, c) I started seeing ads based on the conversation" - b) is never mentioned.
Your points 1-3 are genuinely the best good-faith explanation I've seen of how the "I saw a targeted advert based on something I said with my phone in earshot" thing might happen without it being a deliberate conspiracy between multiple parties.
I still doubt it's actually happening, but I'm not ready to 100% rule out that sequence of events.