• BradleyUffner@lemmy.world
    link
    fedilink
    English
    arrow-up
    17
    arrow-down
    3
    ·
    edit-2
    1 day ago

    The headline seems a little inflammatory. It requires the user to explicitly select a region to share.

    Google is making Gemini a lot more aware of what’s happening inside Chrome. The company has started rolling out a new “Select from screen” feature that lets users highlight specific text or images from a webpage and send them directly to Gemini, making conversations with the AI assistant far more contextual.

    • prole@lemmy.blahaj.zone
      link
      fedilink
      English
      arrow-up
      20
      arrow-down
      3
      ·
      1 day ago

      It’s baffling that over and over and over, every fucking time a corporation does this shit, there’s someone willing to give them the benefit of the doubt.

        • Passerby6497@lemmy.world
          link
          fedilink
          English
          arrow-up
          4
          arrow-down
          4
          ·
          7 hours ago

          “I like to unquestioningly trust people who have repeatedly betrayed my trust in the past, just because they have a token safeguard in place. Surely they won’t rug pull us again and do the thing they always do.”

    • plz1@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      14
      ·
      1 day ago

      The fact that it can do it in the first place means it can do it at all. Just because they require the selection now, doesn’t mean they aren’t harvesting telemetry 24/7, and won’t “accidentally” enable this feature later.

        • Passerby6497@lemmy.world
          link
          fedilink
          English
          arrow-up
          5
          ·
          7 hours ago

          they will not be caught

          They would get caught tho, there’s always some turbo nerd that has a network monitor up and will notice the traffic being sent out. I do it sometimes when I actually have to use chrome, but that’s just my growing distrust of Google.

          Now, whether or not they face repercussions is another story.

    • architect@thelemmy.club
      link
      fedilink
      English
      arrow-up
      5
      arrow-down
      1
      ·
      1 day ago

      No, it can definitely read the entire browser window, fill out and click through forms etc etc. i know, I’ve used it.

      • BradleyUffner@lemmy.world
        link
        fedilink
        English
        arrow-up
        3
        ·
        1 day ago

        That might be true, but it sounds like like a completely different feature than what the article is discussing. This is talking about a new feature that requires the user to select a specific section of the page in order to make it visible to the I’m LLM.

        I am a little confused though, as to why this new feature would be needed if it could already read the entire screen.