In a prior study, the researchers had found that users were increasingly turning to chatbots like ChatGPT to get advice on sensitive topics like health, finance, mental health, and employment, so much so that it was “now becoming more like a social partner than just a tool,” lead author Abhisek Dash said.

Building upon that finding, this time around they also looked into what kinds of sensitive information get stored by the chatbot and just how. They found that 52% of the memory entries contained information about users’ psychological states, including their desires, intentions, emotions, and beliefs.

  • Blaad@europe.pub
    link
    fedilink
    English
    arrow-up
    4
    ·
    6 hours ago

    Leaving this mess behind:

    • 100€ x99 dual Xeon motherboard, 3€ cpus
    • Add in 4 x MI50 32GB and 128GB of ddr4 (~800€ before the hikes of 2025)

    Load up deepseek v4.1 flash q2 and you get opus 4.6 Intelligence at home at good speeds. No snitching.

    Or, if you don’t mind moving slow, just use your brain ig.

  • minorkeys@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    13
    ·
    1 day ago

    Ideally it would be a personal assistant. In reality, it’s a corporate snitch, reporting everything back to the company who will know everything about how to manipulate you into giving up all the power and value of your life.

    • mad_lentil@lemmy.ca
      link
      fedilink
      English
      arrow-up
      5
      ·
      1 day ago

      Yeah then just imagine what happens when the wrong sort of government asks for the keys to this dashboard.

      This technology must be opposed.

      • minorkeys@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        5
        ·
        21 hours ago

        Yeah, what I’ve been saying since the early days of data collection. This is the most extreme version, so far. With neurolink, that’s going to be even more detailed version of knowing our minds. With that one, they don’t need to know each of us as individuals to learn what is common to all of us just from studying the brain data they get from a small group of people.

        We are things to them. Not people. Just things. Just a resource. Like trees or fish or cobalt. Treated no different. This stuff grants more control and deeper value extraction than they already have. These ppl are psychopaths and it’s infuriating watching people empower them further without zero concerns.

        • JustAnotherKay@lemmy.world
          link
          fedilink
          English
          arrow-up
          2
          arrow-down
          1
          ·
          14 hours ago

          Nobody asked an AI to make a portrait, not even in the title. They observed a profile being created by AI agents and used common psychology language to describe the behavior

          • turdburglar@piefed.social
            link
            fedilink
            English
            arrow-up
            1
            ·
            14 hours ago

            you’re missing the point that op was making. it is not a SELF portrait unless is it’s made by the subject. when someone (or something) else makes the portrait, it’s just a portrait.

            you’re also being the tiniest bit belligerent about it.

      • Buffalox@lemmy.world
        link
        fedilink
        English
        arrow-up
        7
        ·
        23 hours ago

        Did you even read the article?

        The AI memory then uses this to build a detailed profile of the user, a mechanism researchers in a study published earlier this year call “the algorithmic self-portrait.”

        Well technically it’s not even a portrait, it’s what is commonly known as a profile, which is also clearly stated first. And it’s definitely not a “self” profile either.
        It’s very Stupid language to use by the researchers.

        • JustAnotherKay@lemmy.world
          link
          fedilink
          English
          arrow-up
          1
          arrow-down
          2
          ·
          14 hours ago

          I would argue that if you’re being pedantic about the language being used, then why are you implying anyone asked an AI to make a portrait of anyone?

          Obviously, yes, it’s not a self-portrait by the dictionary definitions of “self” and “portrait” but people use “self-portrait” to refer to a description of a person all the time, particularly in settings where emotional intelligence is discussed.

          Yes, I’m being pedantic, but so were you, so it feels fair game. But nothing in the article actually makes it seem as if a portrait is being illustrated other than the metaphorical portrait

          • Buffalox@lemmy.world
            link
            fedilink
            English
            arrow-up
            1
            ·
            9 hours ago

            I would argue that if you’re being pedantic about the language being used, then why are you implying anyone asked an AI to make a portrait of anyone?

            I didn’t I just used the semantics of the article and headline. Obviously it’s not a literal “portrait” you’d have to be pretty stupid to read it like that.

            • JustAnotherKay@lemmy.world
              link
              fedilink
              English
              arrow-up
              1
              arrow-down
              1
              ·
              3 hours ago

              Asking an AI to make a portrait…

              This phrase doesn’t imply that someone asked an AI to make a portrait?

              • Buffalox@lemmy.world
                link
                fedilink
                English
                arrow-up
                1
                ·
                18 minutes ago

                Normally it would, but In accordance with the content of the article, the “portrait” is a semaphore.

                Don’t blame me, blame the researchers that thought it was a good idea to use this stupid wording for what is really profiling.

      • mad_lentil@lemmy.ca
        link
        fedilink
        English
        arrow-up
        7
        ·
        1 day ago

        If we can’t be pedantic in the comments of an aggregated story where is there left!

    • Venator@lemmy.nz
      link
      fedilink
      English
      arrow-up
      1
      ·
      7 hours ago

      I guess they’re trying to get at that it’s painted by the user in part…

    • Tehdastehdas@piefed.social
      link
      fedilink
      English
      arrow-up
      5
      ·
      1 day ago

      “Facebook Shadow Profiles”
      https://ideas.repec.org/p/hal/wpaper/hal-04495641.html

      … build shadow profiles by tracking individuals across the web, irrespective of whether they are users … is able to track about 40 percent of the browsing time of both users and non-users … users impose a data externality on non-users by allowing Facebook to infer their personal information.

      • pulsewidth@lemmy.world
        link
        fedilink
        English
        arrow-up
        2
        ·
        8 hours ago

        Yeah, the mechanism discussed in this paper is via Facebook ‘share’ content frames and cookies, not relevant for anyone who disables JavaScript via NoScript or blocks third-party cookies.

        Which I’m guessing based on past comments is 80%+ of Lemmy users. Including me.

  • MalReynolds@piefed.social
    link
    fedilink
    English
    arrow-up
    6
    ·
    1 day ago

    So, surveillance capitalism with extra steps? The steps are your queries and comments? Seems like it’s been done before without as much planet burning.

  • codenul@lemmy.ml
    link
    fedilink
    English
    arrow-up
    8
    ·
    1 day ago

    Is using something like duck.ai doing the same here or does this only affect sign in accounts for the services?

    • scytale@piefed.zip
      link
      fedilink
      English
      arrow-up
      2
      ·
      1 day ago

      Probably not if what they say is true that they anonymize your queries. Not sure if they themselves keep a log though. A VPN and using a container tab or incognito can help you with that, assuming you don’t use a logged in account.

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

      Also interested if Yahoo! Mail will be doing this with their “pay for non-personalised adverts” thing - are they going to be giving all these profiles to advertisers (e.g. foreign intelligence agencies) as part of the “bid for an advertising slot” process? (even if your adblocker then hides that advert)