New warning as AI suddenly targets billions of private messages on smartphones…

Google Update Reveals AI Will Read All Your Private Messages::New warning as AI suddenly targets billions of private messages on smartphones…

  • Ottomateeverything@lemmy.worldEnglish
    721·
    10 months ago

    This is entirely fabricated and opinionated bullshit from Forbes. There is no evidence that this is actually happening and the headline is extremely clickbait. Many other sites have covered this and I can’t find any other than Forbes that take it this way. This is an opinion piece based on a response from a fucking chat bot despite that response semi conflicting with Google’s official outline of the feature.

    Google has announced an AI chat bot function for Bard. You can open a new conversation directly with Bard and in that conversation data is not encrypted and will be sent to Google. There is no evidence of it “reading through your history” or that future chats you have with actual humans are going to be collected or unencrypted.

    If you read past all the “breaks” and such, the author themself even falls back on this:

    For its part, Bard says “Google has assured that all Bard analysis would happen on your device, meaning your messages wouldn’t be sent to any servers. Additionally, you would have complete control over what data Bard analyzes and how it uses it.”

    Let me rephrase that. The author is quoting a fucking response from the BARD CHATBOT and then further extrapolating what the fucking chat bot said. This is not Google. This is A CHAT BOT.

    All of the privacy changes posted by Google themselves do not outline anything more than your conversations with Bard being uploaded. But the author tries to then argue about bards own response he got.

    But I suspect we’ll see that on-device assurance watered down in practice. It will make sense to provide a more seamless interface between a smartphone and the cloud.

    The author is saying he’s scared. And that Google will dance around a loophole its chat bot mentioned. But this is his own opinion about what he thinks will happen because of a response he got from an AI.

    This is fucking ludicrous fear mongering.

    • ParetoOptimalDevEnglish
      41·
      10 months ago

      I really wouldn’t be surprised if Google did this though.

      Plus, wouldn’t it be an easy PR win for them to set the record straight.

      For me these days, Google is guilty until proven innocent.

    • where_am_i@sh.itjust.worksEnglish
      21·
      10 months ago

      Dunno, dude, just that functionality on my phone is fkd up. E.g. “we can now train models in a federated way on a subset of your messages, we promise those are only messages you send to our chatbot and we only extract training results”.

      Nevermind the research showing how to compromise federated training, nevermind privellege escalation, nevermind the backdoor put in for the feds, and so on. I’ma stick to Signal.

  • btaf45@lemmy.worldEnglish
    172·
    10 months ago

    Let’s take advantage of that to train the AI. Put “Convicted Sex Offender Treason Trump needs to be in prison” at the bottom of every message.

  • foggy@lemmy.worldEnglish
    31·
    10 months ago

    I think that’s going to pose big problems for huge swaths of their clients.

    Can any public school lawfully use Google workspace if this is true? I would imagine they can’t.

    • noodlejetski@lemm.eeEnglish
      41·
      10 months ago

      why do you think companies have switched to calling them “direct messages” over the past decade or so?