• Kazumara@feddit.de
    link
    fedilink
    English
    arrow-up
    75
    arrow-down
    2
    ·
    edit-2
    10 months ago

    For really useless call centers this makes sense.

    I have no doubt that a ML chatbot is perfectly capable of being as useless as an untrained human first level supporter with a language barrier.

    And the dude in the article basically admits that’s what his call center was like:

    Suumit Shah never liked his company’s customer service team. His agents gave generic responses to clients’ issues. Faced with difficult problems, they often sounded stumped, he said.

    So evidently good support outcomes were never the goal.

    • mr_tyler_durden@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      ·
      10 months ago

      Agreed. Should we also mourn for the horse and buggy drivers? The gas station attendants? And the whole slew of jobs that have become obsolete over the centuries?

      I do think we need something like UBI and I’m not ignoring the lost jobs but shit jobs shouldn’t have to exist. I’ll mourn for the workers but not for the job. Continuing to employee people to do thankless/hard/dangerous/etc jobs is just silly.

  • uriel238@lemmy.blahaj.zone
    link
    fedilink
    English
    arrow-up
    40
    arrow-down
    2
    ·
    10 months ago

    On one hand, they’re crap jobs. On the other hand, in most economies we have crap jobs not because they’re necessary for productivity, but to give us an excuse to pay people to live.

    Maybe if enough jobs are lost to automation, we’ll start to rethink the structure of a society that only allows people to live if they’re useful to a rich person.

    Essentially, we’re just still doing feudalism with extra steps, and it’s high time we cut that nonsense out.

    • 1984@lemmy.today
      link
      fedilink
      English
      arrow-up
      5
      arrow-down
      2
      ·
      10 months ago

      I think once workers can be replaced, there will be some virus that wipes out most of humanity. No point keeping billions of people around if they aren’t needed.

      • Chaotic Entropy@feddit.uk
        link
        fedilink
        English
        arrow-up
        1
        ·
        10 months ago

        Username checks out… suffice to say that a time of increasing social unrest is on the way, when it’s even easier for the haves to sideline the have nots than it already was.

        • 1984@lemmy.today
          link
          fedilink
          English
          arrow-up
          1
          ·
          10 months ago

          I don’t know, I just think its obvious that the rich guys views ordinary people as useless eaters.

    • Chaotic Entropy@feddit.uk
      link
      fedilink
      English
      arrow-up
      1
      ·
      10 months ago

      We have crappy jobs because jobs need doing and it was still cheaper to get humans to do it without a substantial loss in functionality. They don’t exist because of some form of social altruism, as evidenced by the fact that as soon as a semi-viable alternative is offered then the jobs are gone.

      With the dynamic shifting to automation, prematurely I would add, then employers are seeing a much cheaper way to achieve 80% of what they currently offer.

      • uriel238@lemmy.blahaj.zone
        link
        fedilink
        English
        arrow-up
        1
        ·
        edit-2
        10 months ago

        When I think of crappy jobs I think of a number of different sets.

        Busywork for the extra hands in the clerical pool. This is the stuff that defines the careers of a lot of people in developed countries, in which they’re hired and trained, may even work on projects for a while, and then are dropped into a holding cubical and tasked with sometime benign but probably useless (say entering archived paper files from decades ago into the new data system in case we need them someday – I did that.) Here in the states (and according to anecdotes, the UK) we have a lot of this kind of work, and while it should only be a temporary measure between company projects, entire department clerical pools have been stuck in such holding patterns for years at a time.

        It happens for two reasons I’ve seen: One, the economy tanks such as during the subprime mortgage crisis of 2008, in which lower management in a grand effort of humanitarian desperation, tell the upper management that no, my crew are working hard and very necessary in hopes that theirs is not a department that gets eliminated during the downsizing. (These managers in question are covering their own butts too, but ones I talked with recognized that anyone they dismissed would be eating ramen in a month). And two, the mismanagement of responsibility in linking tasks that need to be done with worker pools capable of doing them. Either the managers tasked to making such links are overwhelmed, or the process of connecting pools to duties is distributed so broadly that it’s de-prioritized by everyone. If those tasks are particularly odious (say they involve interacting with a toxic upper manager) then the lower management will find reasons that their own pool is not able to help, and so the company has simultaneous worker shortages and surpluses. For large, multinational conglomerates, this sort of thing is routine.

        Jobs that are facades to cover for social or moral obligations that are expected of the company, but (from the perspective of shareholders) are too expensive to actually do, such as the faux tech-support services the US exported to phone banks in India that are limited to some very short troubleshooting trees rather than someone actually familiar with the technical aspects of the product. This is (I think) what the business owner of the article is talking about replacing.

        Now what he should be doing is hiring a tech service and including the troubleshooting tree in the manual, what is typically done with household appliances. The workers on that phone bank are being set up with pressure by angry customers to offer some productive solutions, while also getting pressure from management to placate the angry customers, for which they have insufficient facilities. I’m reminded of my own experience being told by upper management I should be spending only fifteen minutes explaining to customers how to install CD-ROM drives (to MS DOS, mind you), which it usually took forty-five minutes to an hour to walk a non-geek through the process.

        Such jobs shouldn’t exist, rather the company should actually hire real departments to deal with social responsibilities, rather than front veneers and marketing campaigns, but that’s a problem intrinsic to the system and not one that will be solved with LLMs given the same short troubleshooting trees. (An LLM with a big troubleshooting tree developed by a serious tech team might work, but would require ongoing development and maintenance, and the occasional tech-support call with a human being. Also a better LLM than we have.)

        Jobs that are odious because they’re labor intensive, hazardous, tedious, frustrating or otherwise taxing on the worker, and yes there are a lot of necessary tasks that need to be done that fall into these categories. So when you say We have crappy jobs because jobs need doing, I assume you’re talking about these.

        Because we’re in a capitalist system that mandates shareholder primacy, our companies first seek out a labor pool they can exploit since they don’t have any other choice. This is classified as bonded servitude, id est slavery but we don’t like to call it that when an enterprise uses human beings like interchangeable, disposable parts. Historically, we’ve hired children, exploited prison populations, immigrants, invoked a truck system, a culture of obligatory productivity, whatever, anything to force our fellow human beings to toil under cruel conditions.

        Without an exploitable population enterprises face labor unrest (unions are the least violent version of this we know) in order to improve conditions and compensation, leaving industries to either capitulate and pay extra and provide proper gear or to automate wherever they can.

        I imagine in collectives, everyone eventually gets pissed off from drawing straws and start working on ways to make odious tasks less odious, either through automation or improving the conditions of the task that it’s no longer odious, e.g. making actual cleaning as close to Power Wash Simulator as possible.

  • realitista@lemm.ee
    link
    fedilink
    English
    arrow-up
    30
    ·
    10 months ago

    I’ve worked in this field for 25 years and don’t think that ChatGPT by itself can handle most workloads, even if it’s trained on them.

    There are usually transactions which must be done and often ad hoc tasks which end up being the most important things because when things break, you aren’t trained for them.

    If you don’t have a feedback loop to solve those issues, your whole business may just break without you knowing.

    • RalphFurley@lemmy.world
      link
      fedilink
      English
      arrow-up
      13
      arrow-down
      2
      ·
      10 months ago

      I ordered Chipotle for delivery and I got the wrong order. I don’t eat meat so it’s not like I could just say whelp, I’m eating this chicken today I guess.

      The only way to report an issue is to chat with their bot. And it is hell. I finally got a voucher for a free entree but what about the delivery fee and the tip back? Impossible.

      I felt like Sisyphus.

      I waited for the transaction to post and disputed the charge on my card and it credited me back.

      There’s so many if-and-or-else scenarios that no amount of scraping the world’s libraries is AI today able to sort out these scenarios.

      • realitista@lemm.ee
        link
        fedilink
        English
        arrow-up
        4
        ·
        10 months ago

        Yes these kind of transactions really need to be hand coded to be handled well. LLM’s are very poorly suited to this kind of thing (though I doubt you were dealing with an LLM at Chipotle just yet).

    • guacupado@lemmy.world
      link
      fedilink
      English
      arrow-up
      2
      ·
      10 months ago

      Maybe you work at a decent place but in my experience you’re really overestimating the people who answer calls and give generic responses.

  • Corhen@lemmy.world
    link
    fedilink
    English
    arrow-up
    28
    ·
    10 months ago

    Seems like a good way to get the “agent” to agree it’s in the wrong, and get 100% refund

    • uriel238@lemmy.blahaj.zone
      link
      fedilink
      English
      arrow-up
      17
      ·
      10 months ago

      I’m interested in if the AI agent has the power to disseminate refunds or at least return authorizations.

      One of the things fascinating to me is that some of the problems humans are bad at handling (such as social engineering) AI tends to be even worse at.

      • Corhen@lemmy.world
        link
        fedilink
        English
        arrow-up
        2
        ·
        10 months ago

        I mean, if you go to your credit card provider with a copy of the log with their rep, and the rep says “i authorize a refund”, you can atleast make the argument.

        Any company scummy enough to trust an AI for this wouldnt give it the authority, though

  • Praise Idleness@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    24
    arrow-down
    3
    ·
    edit-2
    10 months ago
    • works 24/7
    • no emotional damage
    • easy to train
    • cheap as hell
    • concurrent, fast service possible

    This was pretty much the very first thing to be replaced by AI. I’m pretty sure it’d be way nicer experience for the customers.

    • DessertStorms@kbin.social
      link
      fedilink
      arrow-up
      12
      ·
      10 months ago

      I’m pretty sure it’d be way nicer experience for the customers.

      Lmfao, in what universe? As if trained humans reading off a script they’re not allowed to deviate from isn’t frustrating enough, imagine doing that with a bot that doesn’t even understand what frustration is

      • Praise Idleness@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        5
        arrow-down
        4
        ·
        10 months ago

        defacto instant reply, if trained right, way more knowledgeable that the human counterparts, no more support center loop… current experience is such a low bar.

    • gravitas_deficiency@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      4
      arrow-down
      2
      ·
      10 months ago

      Cheap as hell until you flood it with garbage, because there is a dollar amount assigned for every single interaction.

      Also, I’m not confident that ChatGPT would be meaningfully better at handling the edge cases that always make people furious with phone menus these days.

    • philodendron@lemdro.id
      link
      fedilink
      English
      arrow-up
      1
      ·
      10 months ago

      Yeah but are you ready for “my grandma used to tell me $10 off coupon codes as I fell asleep…”

  • 👁️👄👁️@lemm.ee
    link
    fedilink
    English
    arrow-up
    21
    arrow-down
    1
    ·
    10 months ago

    You still need to employ some humans as a backup when the AI catastrophically fucks up, but for the most part it makes sense. Not all jobs need to continue to exist.

    • locuester@lemmy.zip
      link
      fedilink
      English
      arrow-up
      10
      ·
      10 months ago

      Exactly. As the article ends:

      Not every customer service employee should worry about being replaced, but those who simply copy and paste responses are no longer safe, according to Shah.

      “That job is gone,” he said. “100 per cent.”

  • Pratai@lemmy.ca
    link
    fedilink
    English
    arrow-up
    8
    ·
    10 months ago

    Remember when AI was going to make life better for everyone?

    Yeah. That shit’ll be the end of us.

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

    This is just the smallest tip of the iceberg.

    I’ve been working with gpt-4 since the week it came out, and I guarantee you that even if it never became any more advanced, it could already put at least 30% of the white collar workforce out of business.

    The only reason it hasn’t is because companies have barely started to comprehend what it can do.

    Within 5 years the entire world will have been revolutionized by this technology. Jobs will evaporate faster than anyone is talking about.

    If you’re very smart, and you begin to use gpt-4 to write the tools that will replace you, then you MIGHT have 10 good years left in this economy before humans are all but obsolete.

    If you’re not staying up nights, scared shitless by what’s coming, it’s because you don’t really understand what gpt-4 can do.

    • Ann Archy@lemmy.world
      link
      fedilink
      English
      arrow-up
      0
      ·
      10 months ago

      I’m a senior Linux sysadmin who’s been following the evolution of AI over this past year just like you, and just like you I’ve been spending my days and nights tinkering with it non stop, and I have come to more or less the same conclusion as you have.

      The downvotes are from people who haven’t used the AI, and who are still in the Internet 1.0 mindset. How people still don’t get just how revolutionary this technology is, is beyond me. But yeah, in a few years that’ll be evident enough, time will show.

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

          @[email protected]
          @[email protected]
          @[email protected]
          I quite agree.

          And, from SirGolan ref : Submitted on 3 Oct 2023 Language Models Represent Space and Time
          … (from the summary) …Our analysis demonstrates that modern LLMs acquire structured knowledge about fundamental dimensions such as space and time, supporting the view that they learn not merely superficial statistics, but literal world models.
          https://arxiv.org/abs/2310.02207


          What makes it worse (in my opinion) is that LLMs are just one step in this development (which is exponential and not limited by human capabilities).
          For example :
          Numenta launches brain-based NuPIC to make AI processing up to 100 times more efficient
          https://lemmy.world/post/4941919

            • A_A@lemmy.world
              link
              fedilink
              English
              arrow-up
              2
              ·
              5 months ago

              Hi @[email protected],

              since I forgot what I was saying here 4 months ago I read the whole thread again and basically what I said is that I agree with what you said then (4 months ago) and I added a couple of references//ideas to make this point stronger.

              Also, I have no idea why you did receive this notification only today, 4 months after the discussion. I guess the Lemmy software is buggy since for my account I did not receive some notifications in a few instances where someone replied to some of my comments and I just happened to see those replies anyway since I was reading all again.

              take care, 👍