• jws_shadotak@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        27
        ·
        edit-2
        10 months ago

        ChatGPT consistently makes up shit. It’s difficult to tell when something is made up because it’s a language model so it is supposed to sound confident as if it’s any person telling a fact that they know.

        It knows how to talk like a subject matter expert because that’s usually what gets publicized most and thus that’s what it’s trained on, but it doesn’t always know the facts necessary to answer questions. It makes shit up to fill the gap and then presents it intelligently, but it’s wrong.

      • Chahk@beehaw.org
        link
        fedilink
        English
        arrow-up
        15
        ·
        10 months ago

        Most of the time I use assistant to either perform home automation tasks, or look stuff up online. The first one already works fine, and for the second one I won’t trust a glorified autocomplete.

        • stevedidwhat_infosec@infosec.pub
          link
          fedilink
          English
          arrow-up
          2
          ·
          edit-2
          10 months ago

          Good point, hallucinations only add to the fake news problem and artificial content problem.

          I’ll counter with this: how do you know the stuff you look up online is legit? Should we go back to encyclopedias? Who writes those?

          Edit: in case anyone isn’t aware, GPT “hallucinates” made up information in specific cases when temperature and top_p settings aren’t optimized, wasn’t saying anyone’s opinion was a hallucination of course

          • Otter@lemmy.ca
            link
            fedilink
            English
            arrow-up
            5
            ·
            10 months ago

            Some generative chatbots will say something then link to where the info is from. That’s good because I can followup

            Some will just say something. That’s bad and I’ll have to search myself afterwards.

            It’s the equivalent of a book with no cover or a webpage where I can’t see what website it’s on. Maybe it’s reputable, maybe it’s not. Without a source I can’t really decide

            • Sonori@beehaw.org
              link
              fedilink
              English
              arrow-up
              4
              ·
              10 months ago

              Ya, it’s utterly baffling to me that anyone would use a tool that predicts the next word in a sentence to try and learn something. Besides, what’s the endgame when no reporter could make a living because all their words are laundered and fed into a most people are saying bot? At that point new and unknown news, information, and facts will just be filtered out unless a lot of clickbait sites steal them because they the words don’t show up in the average conversation frequently enough.

              Amusing, much like the Cryptocurrency and NFT industry where everyone from the CEO of Openai to the majority of the influencers came from, the extent that the system remind useable at all is reliant on the technology being niche. If it ever actually did become the primary method the tech would fundamentally collapse under its own weight.

            • stevedidwhat_infosec@infosec.pub
              link
              fedilink
              English
              arrow-up
              1
              ·
              edit-2
              10 months ago

              Yep you got me!

              Was leading onto this side of the debate, but basically our collective knowledge, hell our collective experiences are not objective. Our assumptions, mistakes, wordings which result in different interpreted meanings, etc all contribute to some level of disinformation.

              Now let’s not be as nit picky and accept that some detail fudging isn’t the end of the world and happens frequently. We can cross reference each others’ accounts but even that only works to an extent.

              Whole cultures might bare witness to an event and perceive it to be about x y or z, whereas the next door neighbor might see it completely different.

              AI to me really isn’t that far off from the winners being the ones to write the history books, or that strange or unexpected events naturally cause human brains to recollect them in incorrect detail and accuracy.

                • stevedidwhat_infosec@infosec.pub
                  link
                  fedilink
                  English
                  arrow-up
                  0
                  ·
                  edit-2
                  10 months ago

                  Not quite what I meant, I was merely pointing out that we should be cognizant and of how our world view and others views might shape and define what’s considered history or fact.

                  All in all, central points of authority are inherently vulnerable to misinformation. I personally think communal (and biological namely) sources of information shared and verified by each other is far more valuable.

                  Why settle to see the rainbow for your own favorite color when there’s such an amazing and valuable spectrum available. So very digital of us

              • RachelRodent@lemmy.dbzer0.com
                link
                fedilink
                English
                arrow-up
                0
                ·
                edit-2
                10 months ago

                you intentionally refuse to understand what I am saying, yes I don’t use google, samsung or any other assistants either. People talking about me is not a problem corporations and governments spying on me is.

                • stevedidwhat_infosec@infosec.pub
                  link
                  fedilink
                  English
                  arrow-up
                  0
                  ·
                  10 months ago

                  I’m sorry excuse me? I’m not intentionally doing anything. Maybe instead of attributing malice, you might opt for ignorance next time. I can only understand what I understand and you can only communicate as effectively as you can. There’s plenty of room for leeway and benefit of doubt here, unless you’d like another hive mind Reddit clone.

                  Realigning with the conversation, I can understand not wanting powerful parties knowing all about you. They’re much more severe (generally) than people can be. But for some, people can be just as damaging if they have it out for you for whatever reason (gender, sexual orientation, race, success, pissed them off and they’re psychotic, etc)

                  My outlook on privacy isn’t to obscure or hide information but to inflate it with noise instead. Finding a back alley doorway to a building is much easier than finding the right hotel room in a complex of hotels.

                  Obviously not everyone subscribes to this tactic but I wanted to share my outlook as well. I was just sharing what I know and have with you. Wasn’t meant to be a heated debate.