• null@slrpnk.net
    link
    fedilink
    English
    arrow-up
    0
    ·
    2 months ago

    Nope. The amount of data required for constant recording and streaming of audio would be ridiculous and would show up in analysis.

    The simpler answer is that the algorithms are getting scary good.

    • Bizarroland@kbin.social
      link
      fedilink
      arrow-up
      0
      ·
      2 months ago

      You are right, but it is very simple for phones to convert spoken word to text and sending text information back and forth is very lightweight on data usage. So it’s not happening, but it could happen.

    • Wanderer@lemm.ee
      link
      fedilink
      English
      arrow-up
      0
      ·
      edit-2
      2 months ago

      I heard a big one is location. If your mate just came back from Asia and talks to you about it and you then see adds for holidays to Asia. Its because if the location proximity to him, not the conversation. They just assume you would talk about it

      • Switchy85@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        0
        ·
        2 months ago

        Mine does that too. It’s actually all done on device with a special chip and it periodically updates a database on the phone with digital fingerprints for songs. You’ll notice sometimes it’s just wrong or thinks fan noises are music playing.

    • Possibly linux@lemmy.zip
      link
      fedilink
      English
      arrow-up
      0
      ·
      2 months ago

      Not to mention there are a lot of data points that are better. For instance, the can now if your interested in something by monitoring how long you look at it and if you click on something.

    • /home/pineapplelover@lemm.ee
      link
      fedilink
      English
      arrow-up
      0
      ·
      2 months ago

      Download Instagram and start talking about dogs and dog food and it will give you posts and ads about dog food. My friend was talking about random stuff like feet and foot fetishes the other day and you wouldn’t believe what was on their feed when they were scrolling.

    • Ashelyn@lemmy.blahaj.zone
      link
      fedilink
      English
      arrow-up
      0
      ·
      2 months ago

      The device wouldn’t necessarily have to be constantly streaming the audio to a central server. If it’s capable of hearing wake up words like “Ok Google” it’s capable of listening for other phrases and having onboard processing to relay back the results much more compressed. Whether or not this is common practice is another matter, and yes the algorithms are scary good even without eavesdropping.

    • NeoNachtwaechter@lemmy.world
      link
      fedilink
      English
      arrow-up
      0
      ·
      edit-2
      2 months ago

      The amount of data required for constant recording and streaming of audio would be ridiculous

      Relax, it is no big deal. Less than 1 GB for 24 hours of constant recording in good quality.

      Even much less with compression (mp3 etc) and a little loss of audio quality. And if you leave away all the silence, I guess you can tune it down to a few MB.

      That’s one harddisk for a year. But ofc you don’t even want to keep a whole year.