6 ms·

>Idk man, I'm a skeptic that they're not listening in some weird way. Not to mention both of my wife restrict which apps get camera and microphone access. It's

by gruez 1y ago

>Idk man, I'm a skeptic that they're not listening in some weird way. Not to mention both of my wife restrict which apps get camera and microphone access. It's uncanny that things we talk about but never google / look up wind up as ads within a day if not hours. Surely this is easy to test? Come up with a list of 100 topics. Of those, randomly choose 50. Work them into your conversations, and collect all the ads you've seen. Note down how ads you get for the 50 topics you've chosen compared to the topics you haven't chosen. Better yet, give your phone to your friend and have him say the ads, so you don't get confirmation bias. This is all very easy to do, and the conspiracy that facebook/google/whatever is secretly listening to you isn't exactly fringe either. Yet, I'm not of any rigorous testing that proves it's real. While absence of evidence isn't evidence of absence, it's a good reason to be skeptical.

codedokode 1y ago
Let's also consider this theoretically from business point of view. If you were an ad network, and you could listen to what your customers talk about, would you use it or not?
- gruez 1y ago
  If the implication is that ad networks would want to, and therefore it's happening, that's terrible logic. I'm sure the US government would love to convince everyone they landed on the moon and won the space race, without actually having to send a rocket to the moon, but that doesn't mean that means the moon landing was faked.
  - ndriscoll 1y ago
    Part of the point of the space race was to flex on how technologically ahead we were with the implication that our weapons systems were also more advanced, so I suspect that given the choice to actually have more advanced technology vs. merely convincing everyone they did, they would prefer to have actually had that capability. Surveillance companies would presumably also prefer to actually have such capabilities while not broadcasting that fact lest people avoid them or demand they be regulated.
- JohnFen 1y ago
  Well, I'm not a scumbag and have a modicum of ethics so no, I wouldn't even think about using it.
  - mixmastamyk 1y ago
    Hah, you might like this site a friend and I are working on: https://news.ycombinator.com/item?id=45981885 https://news.ycombinator.com/item?id=45981885
grigri907 1y ago
I'm tempted to try this experiment, but I get stuck at "what if my list of 100 things is subconsciously skewed by ads I'm already seeing?" It's the ads I'm seeing, but aren't registering on a conscious level that concern me. I think we see far more ads than we are aware of.
- gruez 1y ago
  That's where the "randomly choose 50" comes in. Even if your initial list of 100 topics was biased, randomly picking from that list allows you to do a comparison between topics that got randomly picked vs ones that didn't. In other words, a randomized controlled trial, the gold standard in experiments. If the group that got randomly picked got a 40% hit rate (ie. that corresponded to ads), but group that didn't randomly get picked got 41%, then you can probably conclude they're not listening to you, even if a 40% hit rate seems spooky. On the other hand if there were significant differences (eg. 40% vs 20%), then there might be other stuff going on worth investigating.
- giancarlostoro 1y ago
  I both love and am horrified by this answer. What if the airport grade tar was there all along? The other wild thing is I was severely out of town, the ad was very local but way too uncomfortably relevant.
strogonoff 1y ago
In 2019, I have done an experiment: talk about a conspicuous topic in presence of an unlocked Google Pixel. Minutes later there were suggestions of relevant to that topic articles in Google app. It is not difficult to believe that if it is available to Google itself, it will be available to whomever that pays for it. That is their business model, after all. The relevant settings are part of your Google account[0], and as of 7 months ago (which is the last time I checked) KB said: > Web & App Activity saves your searches and activity from other Google services in your Google Account. You may get more personalized experiences… > When Web & App Activity is on, you can include audio recordings from your interactions with Google Search, Assistant, and Maps as part of your activity. Speaking in presence of a phone with relevant settings turned on certainly strikes me as something that can qualify as “interaction with Google Assistant”. The KB article does not mention ads, but I would not be surprised if more uses of your voice activity were hidden somewhere in ToS. [0] https://support.google.com/websearch/answer/54068?hl=en&co=GENIE.Platform%3DAndroid#zippy=%2Caudio-recordings https://support.google.com/websearch/answer/54068?hl=en&co=G...
- gruez 1y ago
  >In 2019, I have done an experiment: talk about a conspicuous topic in presence of an unlocked Google Pixel. Minutes later there were suggestions of relevant to that topic articles in Google app. That's hardly a conclusive experiment when there wasn't a control group, and sample size was 1. That's the issue with the "facebook/google/whatever is secretly listening to you" conspiracy theories: plenty of anecdotes, zero rigorous testing.
  - strogonoff 1y ago
    Sample size is small but you see a KB article clarifying that it is what happens. It is not “secretly listening” if they say it openly. I never said “secretly” either.
    - gruez 1y ago
      >small but you see a KB article clarifying that it is what happens. "and activity from other Google services" and "audio recordings from your interactions with Google" is hardly conclusive, especially when google has services that involve consensual audio recordings.

by gruez 1y ago

codedokode 1y ago

Let's also consider this theoretically from business point of view. If you were an ad network, and you could listen to what your customers talk about, would you use it or not?

gruez 1y ago
If the implication is that ad networks would want to, and therefore it's happening, that's terrible logic. I'm sure the US government would love to convince everyone they landed on the moon and won the space race, without actually having to send a rocket to the moon, but that doesn't mean that means the moon landing was faked.
- ndriscoll 1y ago
  Part of the point of the space race was to flex on how technologically ahead we were with the implication that our weapons systems were also more advanced, so I suspect that given the choice to actually have more advanced technology vs. merely convincing everyone they did, they would prefer to have actually had that capability. Surveillance companies would presumably also prefer to actually have such capabilities while not broadcasting that fact lest people avoid them or demand they be regulated.
JohnFen 1y ago
Well, I'm not a scumbag and have a modicum of ethics so no, I wouldn't even think about using it.
- mixmastamyk 1y ago
  Hah, you might like this site a friend and I are working on: https://news.ycombinator.com/item?id=45981885 https://news.ycombinator.com/item?id=45981885

grigri907 1y ago

I'm tempted to try this experiment, but I get stuck at "what if my list of 100 things is subconsciously skewed by ads I'm already seeing?" It's the ads I'm seeing, but aren't registering on a conscious level that concern me. I think we see far more ads than we are aware of.

gruez 1y ago
That's where the "randomly choose 50" comes in. Even if your initial list of 100 topics was biased, randomly picking from that list allows you to do a comparison between topics that got randomly picked vs ones that didn't. In other words, a randomized controlled trial, the gold standard in experiments. If the group that got randomly picked got a 40% hit rate (ie. that corresponded to ads), but group that didn't randomly get picked got 41%, then you can probably conclude they're not listening to you, even if a 40% hit rate seems spooky. On the other hand if there were significant differences (eg. 40% vs 20%), then there might be other stuff going on worth investigating.
giancarlostoro 1y ago
I both love and am horrified by this answer. What if the airport grade tar was there all along? The other wild thing is I was severely out of town, the ad was very local but way too uncomfortably relevant.

strogonoff 1y ago

In 2019, I have done an experiment: talk about a conspicuous topic in presence of an unlocked Google Pixel. Minutes later there were suggestions of relevant to that topic articles in Google app. It is not difficult to believe that if it is available to Google itself, it will be available to whomever that pays for it. That is their business model, after all. The relevant settings are part of your Google account[0], and as of 7 months ago (which is the last time I checked) KB said: > Web & App Activity saves your searches and activity from other Google services in your Google Account. You may get more personalized experiences… > When Web & App Activity is on, you can include audio recordings from your interactions with Google Search, Assistant, and Maps as part of your activity. Speaking in presence of a phone with relevant settings turned on certainly strikes me as something that can qualify as “interaction with Google Assistant”. The KB article does not mention ads, but I would not be surprised if more uses of your voice activity were hidden somewhere in ToS. [0] https://support.google.com/websearch/answer/54068?hl=en&co=GENIE.Platform%3DAndroid#zippy=%2Caudio-recordings https://support.google.com/websearch/answer/54068?hl=en&co=G...

gruez 1y ago
>In 2019, I have done an experiment: talk about a conspicuous topic in presence of an unlocked Google Pixel. Minutes later there were suggestions of relevant to that topic articles in Google app. That's hardly a conclusive experiment when there wasn't a control group, and sample size was 1. That's the issue with the "facebook/google/whatever is secretly listening to you" conspiracy theories: plenty of anecdotes, zero rigorous testing.
- strogonoff 1y ago
  Sample size is small but you see a KB article clarifying that it is what happens. It is not “secretly listening” if they say it openly. I never said “secretly” either.
  - gruez 1y ago
    >small but you see a KB article clarifying that it is what happens. "and activity from other Google services" and "audio recordings from your interactions with Google" is hardly conclusive, especially when google has services that involve consensual audio recordings.