Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

I still don't think companies serve you ads based on your microphone

simonwillison.net

Post reply on HN
Print view View original

251–260 of 361 posts

Re: I still don't think companies serve you ads based on your microphone

#251

Post by BriggyDwiggs42 »

MrMcCall wrote: post #244
BriggyDwiggs42 wrote: post #225

Earlier quoted context omitted.

Im not the guy, but you never bother trying to prove anything; you just ramble about vibes most people will implicitly agree with. “Oh boy rich people sure do suck and companies sure do suck.” “Aww geez my wife got a targeted ad.” I loosely figure that they spy too, but I expected a better showing from HN in rebuttal to the article. If it’s so damn obvious then show some actual proof.

I'll get right on that. Wait right here.

Hey if you dig something up I’m happy to hear it. You’d be confirming my suspicions.

Re: I still don't think companies serve you ads based on your microphone

#253

Post by FrustratedMonky »

snowwrestler wrote: post #99
alvah wrote: post #29

Every time I read some technical description about why this isn't happening, the technical description seems convincing. However... A friend tested the theory a few years ago. He doesn't own a swimming pool, doesn't want to, and has never expressed any desire to. He put his and his wife's phone on the table and said to the wife (loudly), "Why don't we look into pool fencing?". She agreed with him. Shortly after, on b…

Think about what this implies. If your phone is listening, it’s listening all the time, right? So like 12-18 hours of continuous audio every day. That’s a lot of ad triggers. Way too many to actually be served with ads during your browsing time, which is a strict subset of your total audible proximity to your phone (plus ad inventory is a strict subset of what you view on your phone). So how does the phone + ad netwo…

"How would it do that? "

AI.

That is entire premise of 'Nexis' from Yuval Harari.

Individualized bot driven surveillance .

https://www.theregister.com/2024/09/16/oracle_ai_mass_survei...

""Ellison declares Oracle all-in on AI mass surveillance, says it'll keep everyone in line

Cops to citizens will be 'on their best behavior because we're constantly recording and reporting'""

Re: I still don't think companies serve you ads based on your microphone

#254

Post by marcusverus »

snowwrestler wrote: post #99
alvah wrote: post #29

Every time I read some technical description about why this isn't happening, the technical description seems convincing. However... A friend tested the theory a few years ago. He doesn't own a swimming pool, doesn't want to, and has never expressed any desire to. He put his and his wife's phone on the table and said to the wife (loudly), "Why don't we look into pool fencing?". She agreed with him. Shortly after, on b…

Think about what this implies. If your phone is listening, it’s listening all the time, right? So like 12-18 hours of continuous audio every day. That’s a lot of ad triggers. Way too many to actually be served with ads during your browsing time, which is a strict subset of your total audible proximity to your phone (plus ad inventory is a strict subset of what you view on your phone). So how does the phone + ad netwo…

1) App stores the trailing two minutes of speech in memory.

2) If the app detects a consumption-related trigger word, the related conversation is flagged for transmission to the server.

3) Flagged audio block is converted to text. Consumption related verbs ("buy", "purchase", etc) are identified. The syntax of the sentence clearly indicates which noun is the target of a given consumption-related verb ("new car", "pool fencing")

4) Serve related ads

Re: I still don't think companies serve you ads based on your microphone

#255

Post by simonw »

marcusverus wrote: post #254
snowwrestler wrote: post #99

Earlier quoted context omitted.

Think about what this implies. If your phone is listening, it’s listening all the time, right? So like 12-18 hours of continuous audio every day. That’s a lot of ad triggers. Way too many to actually be served with ads during your browsing time, which is a strict subset of your total audible proximity to your phone (plus ad inventory is a strict subset of what you view on your phone). So how does the phone + ad netwo…

1) App stores the trailing two minutes of speech in memory. 2) If the app detects a consumption-related trigger word, the related conversation is flagged for transmission to the server. 3) Flagged audio block is converted to text. Consumption related verbs ("buy", "purchase", etc) are identified. The syntax of the sentence clearly indicates which noun is the target of a given consumption-related verb ("new car", "poo…

Where's the proof that this is happening?

Lots of people run network traffic sniffers to see what apps are doing. Lots of people decompile apps. Lots of people at companies leak details of bad things they are doing.

Why has nobody been able to demonstrate this beyond anecdotes about talking about swimming pools and then getting adverts for swimming pool stuff?

Re: I still don't think companies serve you ads based on your microphone

#256

Post by simonw »

FrustratedMonky wrote: post #253
snowwrestler wrote: post #99

Earlier quoted context omitted.

Think about what this implies. If your phone is listening, it’s listening all the time, right? So like 12-18 hours of continuous audio every day. That’s a lot of ad triggers. Way too many to actually be served with ads during your browsing time, which is a strict subset of your total audible proximity to your phone (plus ad inventory is a strict subset of what you view on your phone). So how does the phone + ad netwo…

"How would it do that? " AI. That is entire premise of 'Nexis' from Yuval Harari. Individualized bot driven surveillance . https://www.theregister.com/2024/09/16/oracle_ai_mass_survei... ""Ellison declares Oracle all-in on AI mass surveillance, says it'll keep everyone in line Cops to citizens will be 'on their best behavior because we're constantly recording and reporting'""

How were they doing this in 2017?

Re: I still don't think companies serve you ads based on your microphone

#257

Post by alpaca128 »

alvah wrote: post #29

Every time I read some technical description about why this isn't happening, the technical description seems convincing. However... A friend tested the theory a few years ago. He doesn't own a swimming pool, doesn't want to, and has never expressed any desire to. He put his and his wife's phone on the table and said to the wife (loudly), "Why don't we look into pool fencing?". She agreed with him. Shortly after, on b…

It is interesting how people always come up with anecdotes like this but none of them try repeating the experiment multiple times.

You might think the pool fencing example might be an extreme coincidence, but far weirder things happen every day. And what made your friend consider pool fencing as an example if they don't like pools? Maybe something they saw recently gave them the idea? Hmm...

Re: I still don't think companies serve you ads based on your microphone

#258

Post by _1tem »

snowwrestler wrote: post #117

This doesn’t seem that hard to test? Put a phone on WiFi, run it through a traffic sniffer, and see if it is constantly sending little audio packets to listener1486.facebook.com or whatever. Or heck, open up a phone and stick a probe on the mic lead. See when it is getting power or not.

There are smarter ways to do this without constantly sending all the audio. Wakewords + surrounding snippets for example. In fact if the audio recognition is done on-device then only the text data could be sent which would be indistinguishable from telemetry.

Re: I still don't think companies serve you ads based on your microphone

#259

Post by marcusverus »

simonw wrote: post #255
marcusverus wrote: post #254

Earlier quoted context omitted.

1) App stores the trailing two minutes of speech in memory. 2) If the app detects a consumption-related trigger word, the related conversation is flagged for transmission to the server. 3) Flagged audio block is converted to text. Consumption related verbs ("buy", "purchase", etc) are identified. The syntax of the sentence clearly indicates which noun is the target of a given consumption-related verb ("new car", "poo…

Where's the proof that this is happening? Lots of people run network traffic sniffers to see what apps are doing. Lots of people decompile apps. Lots of people at companies leak details of bad things they are doing. Why has nobody been able to demonstrate this beyond anecdotes about talking about swimming pools and then getting adverts for swimming pool stuff?

These are fair questions! I'm not convinced that it is happening. Nor am I convinced, as the parent seems to be, that it would be difficult to do.

edit: Having re-read my comment, I can see how it could easily be read to say "It's happening and this is how it works", whereas I intended to convey something like "It could easily be done and here's how." I have a bad habit of implying my point rather than stating it outright. I'm working on it!

Re: I still don't think companies serve you ads based on your microphone

#260

Post by ComplexSystems »

simonw wrote: post #207
siliconpotato wrote: post #188

Earlier quoted context omitted.

why are you limiting your argument to wake words? It's comparatively trivial to serve ads based upon what's spoken after the wake word. And as we know from the lawsuit, it seems that there's been a lot of data gathered accidentally too.

The story here is about Apple paying a settlement because Siri was sending audio snippets surrounding invalid wake word activations back to their servers without users realizing it. In the story, there was an implication that these audio snippets had been used for targeted ads. This is very clearly not true. A separate issue is whether Apple take audio from opt-in full hey Siri sessions and sell that advertisers for…

None of this is in the realm of "conspiracy theory." We aren't talking about the Earth being flat. Your phone is very clearly listening at all times, and sometimes activates and sends that data to servers. Given that, we can debate all day about what happens with that data - hopefully it's being treated properly - but the point is that it isn't clear, there isn't enough transparency, and there are occasionally scandals.

The problem is that the infrastructure to harvest all of this data clearly exists, and the only reason we shouldn't believe that it's being mishandled is whatever degree of faith we have in these companies to behave ethically, comply with legal requirements (or face a slap on the wrist), and most importantly, have developers that don't inadvertently make any mistakes. I really don't have faith in the latter point in particular.

The point is that maybe this data improperly ends up being used to serve ads, or maybe it doesn't, but in light of all the above, entertaining the idea is in no way akin to thinking that the moon is made of cheese.

EDIT: clarified "recording" to "listening".

Post reply on HN