Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror
×
AI Programming IT

Survey Claims Some Companies are Already Replacing Workers With ChatGPT (yahoo.com) 142

An anonymous reader quotes an article from Fortune: Earlier this month, job advice platform Resumebuilder.com surveyed 1,000 business leaders who either use or plan to use ChatGPT. It found that nearly half of their companies have implemented the chatbot. And roughly half of this cohort say ChatGPT has already replaced workers at their companies....

Business leaders already using ChatGPT told ResumeBuilders.com their companies already use ChatGPT for a variety of reasons, including 66% for writing code, 58% for copywriting and content creation, 57% for customer support, and 52% for meeting summaries and other documents. In the hiring process, 77% of companies using ChatGPT say they use it to help write job descriptions, 66% to draft interview requisitions, and 65% to respond to applications.

Overall, most business leaders are impressed by ChatGPT's work," ResumeBuilder.com wrote in a news release. "Fifty-five percent say the quality of work produced by ChatGPT is 'excellent,' while 34% say it's 'very good....'" Nearly all of the companies using ChatGPT said they've saved money using the tool, with 48% saying they've saved more than $50,000 and 11% saying they've saved more than $100,000....

Of the companies ResumeBuilder.com identified as businesses using the chatbot, 93% say they plan to expand their use of ChatGPT, and 90% of executives say ChatGPT experience is beneficial for job seekers — if it hasn't already replaced their jobs.

This discussion has been archived. No new comments can be posted.

Survey Claims Some Companies are Already Replacing Workers With ChatGPT

Comments Filter:
  • by thesjaakspoiler ( 4782965 ) on Sunday February 26, 2023 @07:17PM (#63325350)

    Is it OK to submit a resume written by ChatGPT?

  • Seems legit. (Score:5, Insightful)

    by Timtheenchanted ( 899695 ) on Sunday February 26, 2023 @07:28PM (#63325378)
    One source of meaningless work salad is probably as good as another. THe bot is probably an ideal replacement for most executive roles.
  • It will be interesting to see what happens when ChatGPT submits misleading or outright false public filings or regulatory information -- who will be held responsible?

    • ChatGPT doesn't submit anything. An employee representing the company as a legal entity does & so the company is liable. ChatGPT would be no more responsible than a typewriter. I'm guessing many clerical jobs will become proof-reading & fact-checking LLM output.
      • by NFN_NLN ( 633283 )

        Is Congress liable for voting on a 40,0000 page bill after having it for only 24 hours? Ie. not enough time for a human to even flip through all the pages.

        • Is Congress liable for voting on a 40,0000 page bill after having it for only 24 hours? Ie. not enough time for a human to even flip through all the pages.

          Ironically you've made a justification to replace those we elect to Congress with scripts more than anything.

          It's rather clear they don't intend to do their jobs to any human level when presenting the impossible on lawmakers, and the only reason to do so, is to deceive.

          • You misunderstanding what their jobs are. They are there to suck up money, enrich themselves and make laws that help whoever donates the most to their reelection. It's clearly working as intended.

            Oh, you thought government was for the people? Grow up.

        • Is Congress liable for voting on a 40,0000 page bill after having it for only 24 hours? Ie. not enough time for a human to even flip through all the pages.

          Yes.

          Good luck fixing that though.

          • ChatGPT is pretty consistently good at summarising submitted texts though. May reduce those pages down to human readable point form? But yeah, the point is to create obfuscated laws that benefit their authors, e.g. ALEC, & that only corporate lawyers can make any useful sense of. The submitted bills often still have the placeholder text & document metadata that shows who actually drafted them, i.e. not anyone who was elected or works as a public servant.
    • It will be interesting to see what happens when ChatGPT submits misleading or outright false public filings or regulatory information -- who will be held responsible?

      (Every Lawyer) "Who the fuck cares. All I see, is dollar signs."

    • Re:Who is liable? (Score:5, Informative)

      by slack_justyb ( 862874 ) on Sunday February 26, 2023 @08:50PM (#63325578)

      who will be held responsible?

      Why do people keep asking these kinds of questions? We have automation tools in the legal field that do all kinds of rote work. If it messes up, it's the law office that's on the hook for "should of caught it". This kind of question is already answered. If you use automation and it messes up, it's still on you to deal with the fall out. If you then really want to, you can go after the person who made the tool, but only after you've taken your hit. The only way you get to pass the buck on blame is if Congress or your State assembly passed a law letting you do so. And right now the majority of passing the buck laws we've got only deal with if you contract out. And even then only if your contract had language in it to allow so.

      So long story short, if your automation messes up, you're the one that's got to deal with it. If some public filings mess up, it's the company that the filings apply to and more specifically the CEO's ass that's on the line (which I say that, but the biggest oh no that'll happen is they get fined which then gets rolled into the company).

      • it's the law office that's on the hook for "should of caught it".

        Let's hope they at least know how to spell "should have"...

    • It will be interesting to see what happens when ChatGPT submits misleading or outright false public filings or regulatory information -- who will be held responsible?

      With the existence of "corporation" abusive loopholes and laws, I'm just curious; why do you assume corporations today would suddenly care about responsibility? They certainly haven't proven it before when they were employing humans instead of scripts.

    • Boss: Your jobs is being replaced by ChatGPT, sorry.
      Grunt: So I'm being fired?
      Boss: No, just reassigned. Your new job is to type questions into ChatGPT instead of thinking of the answers yourself.
      Grunt: Well, I guess...
      Boss: and also a 50% pay cut.

  • Remember when "Go away or I will replace you with a very small shell script." used to be a joke?

    Next thing you know, ChatGPT will escape and start eating lawyers.

    • by Dutch Gun ( 899105 ) on Sunday February 26, 2023 @08:02PM (#63325450)

      See, the humor in that statement was that there are actually a lot of people whose jobs could long have been automated away with some fairly basic tools and logic. But not everyone is good at writing those tools, so the job stays safe.

      Now that we have a method of more easily automating fairly simple, repetitive tasks, or producing code snippets on demand, anyone's job that involved such things are likely in trouble, and will have to either expand their capabilities or find new work.

      But still, most programmers, that is, people who actually solve real problems and don't just copy / paste / tweak shit are still safe for the time being. I think AI can get part of the way there as a great tool for helping programmers, but being able to completely understand in depth how to deal with an actual production environment seems like it's going to require near human-level understanding. At the moment, what ChatGPT can do is impressive, but when you examine it more closely, it's incredibly broad in capabilities, but still quite shallow.

      • ChatGPT is essentially a research assistant. It is quite useful for that, especially within realms that it is well trained on. Still needs an expert to assess the information though and to some degree write the prompts.

      • In many ways I think that ChatGPT will be *better* than a large number of the programmers out there. Converting a deep understanding of a problem into code is hardly a difficult task. Reading existing code from a functional perspective isn't really that hard either. The high-value work is being able to think through the actual problem you are trying to solve. Most of the interesting software doesn't solve a problem with a physical model but rather tries to codify organically evolved human behavior. The
    • Next thing you know, ChatGPT will escape and start eating lawyers.

      You act like that's some kind of threat to those who have literally written the legal system by lawyers, for lawyers.

      There's a reason we don't call it a "Justice" system anymore. That said, you might be thanking a lawyer in the future, human. They might just stand as a fellow meatsack and defend your worth by justifying your continued employment against replacing you with a shell script.

      (Yeah, I hate lawyers too. Doesn't mean we won't need them in the future, just like you need them when you need them n

      • The main reason you need them is the law is written in such a way that you need them. We need to simplify the legal system so the person with the biggest wallet doesn't usually win.

        • The main reason you need them is the law is written in such a way that you need them.

          Let me know when a legal system written by them for them didn't convince you of that delusion. Then try and remember the concept of justice has existed LONG before the concept of "needing" a lawyer has.

          We need to simplify the legal system so the person with the biggest wallet doesn't usually win.

          Fucking hell. Talk about delusions. PROVE that the biggest wallet doesn't usually win first. THEN you can stand on that fucking delusion as a defense.

        • There is no way to simplify the legal system such that lawyers will not be needed to navigate it. As much as we would like to believe that lawyers are only necessary because lawyers have forced their own necessity, it simply isn't true. They have never needed to do so. The enterprise of legal regulation is naturally complex enough that you will always need specialists to understand it.

          Realize that we are talking about a system of regulation that manages hundreds of millions of people, most of whom are th

    • Damn you all to helllllllll!n!^111
    • > Next thing you know, ChatGPT will escape and start eating lawyers.

      I believe that's already started. https://www.reuters.com/legal/... [reuters.com]

  • by Ol Olsoc ( 1175323 ) on Sunday February 26, 2023 @07:42PM (#63325412)
    In a world where the next quarter determines success, what happens then? If you eliminate the employees, and profit immensely, you have to increase that profit the next cycle.

    It's an interesting mind game. Which do we reach first, the point where there aren't enough people to purchase products because they are no longer employable, or the point where the executives lose their jobs because their effort is replaceable?

    • Could work if prices were to reduce drastically in sync with people making much less (I am reminded of the head of Panasonic saying something to the effect of making everything as cheap/accessible as water. Then you have opened your potential customers immensely). You could get to the worries of 60s economist of everyone working part time to maintain a comfortable lifestyle.

      But

      The move now seems to be buy the earth outright and turn everyone into a sharecropper.

      • Reducing prices doesn't fuel profits very much unless you can make up for it with higher margins and greater volume. In a competitive market, sure, you can do that. But once all labor costs are reduced, you won't really be able to undercut the competition much and there won't be a market.
  • by 93 Escort Wagon ( 326346 ) on Sunday February 26, 2023 @07:46PM (#63325420)

    All these ChatGPT stories are being driven by the companies and investors behind ChatGPT, who intend to do an IPO in the next 6-12 months.

    • All these ChatGPT stories are being driven by the companies and investors behind ChatGPT, who intend to do an IPO in the next 6-12 months.

      Snap Inc. filed an IPO where a reclaimed valuation of their company was reduced to "only" $20 billion. In their IPO filing documents, they claimed they had never made a profit, and may never make a profit. THAT, is how fucked "investing" is today.

      If someone wishes to carry that delusion into a recession, at least let me grab my popcorn. This should be fucking rich.

      • Ever read an S1 filing? They list everything from solar flares to nuclear war as risks. The list of risks in an S1 is stuffed, somewhat intentionally, with everything they can possibly think of in the case that it happens. Public companies are sued, constantly, for general day to day operations because some investor somewhere will scream you didn't tell me that the collapse of society will negatively affect my investment.

        I'm not saying any of these things are good nor bad, really - but pointing to someon

    • by VeryFluffyBunny ( 5037285 ) on Sunday February 26, 2023 @08:45PM (#63325562)
      Yep. Remember when Google were supposed to replace doctors & epidemiologists with their predictive algorithms? Their claims turned out to be mostly hot air & their super-smart tools only useful for speeding up specific processes under specific conditions. Once the ChatGPT hype has calmed down, people have found out its shortcomings & decided it's no good for anything, & then after a while some people actually find some practical, real world uses for it, then we'll see what it's actually good for & what it's worth.

      On the plus side, overly confident bullshit, like output from ChatGPT, does seem to be a remarkably marketable commodity these days. If a machine can churn it out, that's a lot cheaper than sitting humans at keyboards to generate bullshit all day. The thing is, once people realise what this stuff is, it it loses it's appeal, will they want to read it anymore? Will there be a human audience for it?

      Then there's the issue of copyright but not how it's typically being argued at the moment. The thing is, courts have already decided on multiple occasions that output from algorithms, robots, machines, etc., i.e. produced without human intentional control, can't be copyright licensed. That means that all of ChatGPT's output is effectively public domain & anyone can do anything they want with it. It'll be interesting to see what this does, for example, to the publishing industry.
      • Yep. Remember when Google were supposed to replace doctors & epidemiologists with their predictive algorithms? Their claims turned out to be mostly hot air & their super-smart tools only useful for speeding up specific processes under specific conditions.

        Uh, when the doctor drug pimp slinging sponsored answers is the "solution" we have today, don't pretend even an artificial intelligence cannot discern between a legitimate business and massive corruption.

        It's not exactly hard to see the cracks in the system today. Human greed is involved. Go fucking figure.

      • For many things I have just as much luck with Google as I do the doctor. And Google lets us go to the doctor better prepared. Much like ChatGPT won't be a complete replacement it will change how things get done.
        • You think your few hours of Googling is anywhere near equivalent to your doctor's expertise? Do you think your doctor could spend a few hours on Google & do your job better than you?
          • The doctor is better but they cost significantly more. If you can save the doctor time, you can save your money. You are not replacing the doctor, just making things more efficient. Doctor's already barely spend any time with patients anyway.

            • Exactly. I get like 15 minutes with a doctor to tell them my symptoms and whatnot. If I can learn the correct terms and be efficient, even that is a win over just showing up.
            • Sounds like you have a consumer-driven healthcare system. Over here, doctors will typically listen to a patient recount what they read on Google, probably disregard it, & then talk to the patient about their symptoms, maybe do an examination, maybe some tests, & then try to diagnose whatever the evidence seems to point to. They may not get the diagnosis, prognosis, & treatment right first time because patients usually present with a wide range of symptoms with a wide range of possible causes
      • by mrfaithful ( 1212510 ) on Monday February 27, 2023 @09:10AM (#63326554)

        If a machine can churn it out, that's a lot cheaper than sitting humans at keyboards to generate bullshit all day.

        Buzzfeed had an army of low paid humans to generate clickbait articles based on some very simple rules. This was determined by data that "Top X lists where number N will surprise you!" and "which power ranger are you?" did very well. And what happened? They refined their data driven process to the point they made themselves samey, irrelevant and just part of the background noise.

        ChatGPT may be able to optimise this race to irrelevance, but I just don't see how a program doing statistical modelling and prediction can somehow achieve anything beyond what we already have but worse?

    • There's probably some shilling since Microsoft is involved, but I also think a lot of people (myself included) are just legitimately excited about it. ChatGPT has already replaced Google as my first step in looking things up; Google is now the second stop in most cases if ChatGPT's answer seems suspicious.
  • Regurgitrant is the sum of all our posts. Just like search results always have been. We're just using that to a new degree. If your job can be replaced by regurgitrant then you were already redundant when you were hired. No sympathy here. Anyway, there will be new jobs writing prompts for AI, go get one of those.
  • by mark-t ( 151149 )

    At its core, a ChatGPT is simply designed to predict the next word or sequence of words in a piece of text based on the preceding context. This means that it has no actual understanding of what it is being told or what it is producing as output. The fact that it often appears to be coherent is more of a result of projection by the user than any reflection of its true capabilities.

    The only jobs that ChatGPT could possible really replace as is are jobs where the accuracy of any output does not actually m

    • Re:How? (Score:5, Interesting)

      by godrik ( 1287354 ) on Sunday February 26, 2023 @09:02PM (#63325608)

      I don't think that chatGPT is replacing a worker as much as chatGPT makes good employees 50% more efficient and so you can lay off 30% of your work force

    • The only jobs that ChatGPT could possible really replace as is are jobs where the accuracy of any output does not actually matter, and all that counts is that the output "look" reasonably convincing to non-experts.

      So ... first level tech support and politics.

    • Very easily. Accuracy is no problem at all, at the end of the day, there is still a human checking the work over before it's done. Most jobs involve a lot of fluff and boilerplate that isn't really novel work or add anything substantial, but it still needs to be done and it still costs money. That's what ChatGPT and the likes automate away. So it's a great efficiency improvement, let the machine do the monkey work and focus on what actually matters.
  • by PPH ( 736903 ) on Sunday February 26, 2023 @08:49PM (#63325576)

    I had a co-worker who took the old 'spew' program (which wrote faux headlines from a vocabulary/grammar file and an RNG), reloaded the vocabulary file with standard corporate-speak and put it to work writing corporate mission/vision statements.

    We had a contest where real mission/vision statements (suitably anonymized) were mixed in with the output from spew. It was surprisingly difficult to pick out the actual source for each.

  • Forget your database going down and halting all your business, at least it is possible to recover from that. Replacing workers with ChatGPT truly means that your business now beholden to one company. More so, they can and will skew what ChatGPT tells to advantage themselves.
    • Forget your database going down and halting all your business, at least it is possible to recover from that. Replacing workers with ChatGPT truly means that your business now beholden to one company. More so, they can and will skew what ChatGPT tells to advantage themselves.

      With all the evil that Larry Lawyer Inc. has brought upon business by selling a database product instead of a licensing lawsuit, I cannot imagine business being dumb enough to sign up for an Oracle ChatGPT.

      Or at least as long as we hire corrupt greedy humans as CEOs.

  • Survey says... (Score:5, Insightful)

    by Gimric ( 110667 ) on Sunday February 26, 2023 @09:15PM (#63325616)

    So a survey of "1,000 business leaders who either use or plan to use ChatGPT" found that almost half of them have implemented ChatGPT?

    Amazing insight. Did the other half plan to implement ChatGPT?

  • by CmdrPorno ( 115048 ) on Sunday February 26, 2023 @10:21PM (#63325724)

    and nobody noticed!

  • ChatGPT is a non-thinking non-expert and crap programmer.
    • Put it into management, it seems qualified.

    • A lot of jobs don't require thinking. They require you to do repeatable tasks. Plenty of jobs will still exist, but you will need fewer of them to accomplish more in the same set of time.

  • by Fly Swatter ( 30498 ) on Sunday February 26, 2023 @10:58PM (#63325780) Homepage
    ChatGPT burst onto the scene what, a few months ago? And already replaced half the employees at some random .com website? Wow. This is amazing!

    I will be convinced it has actually become useful when the Indian call center jobs are all gone. It's the lowest hanging fruit.
  • Unofficially, third world call centers are using this for mundane calls, and for calls that face legal or attorney prosecution - such as demanding ID. I know one lawyer who earns $120K a year or about 9K per unnecessary dispute, serving board members personally. Amazing how many financial institutions do not know the law(s) and try to leave you in call centre hell. Never tell a lawyer 'these are our rules' after he verbals the actual law. AI is adapting however- not replying to people on their 'hotlist of l
  • I can only really see one good use of chat GPT and similar markov-chain-generators: undoing their own work.

    When I see a pile of word salads and I dont want to bother reading it, if an un-chat-gpt bot could tell me "what inputs to you would result in this output" it would save me a lot of time.

    "Create a legal boilerplate for XYZcorp"

    "Make an argument in favor of idea ABC."

    "What are the best criticisms of EFG?"

    etc, will tell me what im looking at without having to waste time eyeballing all the nonsense.

    Even b

    • by narcc ( 412956 )

      Just a nit-pick: The 'T' in 'ChatGPT' stands for 'transformer'. Those are decidedly not Markovian.

  • Start Laying the groundwork to start exploring exoplanets: https://aeon.co/videos/burning... [aeon.co] But no profit in it, oh well, can't be done then.
  • Business leaders already using ChatGPT told ResumeBuilders.com their companies already use ChatGPT for a variety of reasons, including 66% for writing code, 58% for copywriting and content creation, 57% for customer support, and 52% for meeting summaries and other documents.

    ... and 100% for responding to questions about what ChatGPT is used for.

    Hmm, companies are using ChatGPT for meetings that just happened, i.e., after ChatGPT was trained. Maybe I can get ChatGPT to write the summaries now of my meetings for the next year. Hello extra quiet quitting time!

    • Having ChatGPT write a summary of a meeting seems like a great idea. It doesn't go into the meeting with preconceived notions. And if the summary comes out weird, you know that maybe you didn't state all of the implicit assumptions and should go back for another round of clarification.
  • "Oh, sorry, that was Al, short for 'Always Learning', our new AI, who had you in tears. Can I put you on hold while I find a tech to talk to you, or would you like to call back .... Hello? Oh darn, they hung up. Mwuhahaha!"

  • 89% satisfaction? That's about 89% higher than with human management, the solution is so obvious! ChatGPT did pass a MBA master exam and it's FREE, compared to some of the most expensive employees of a company.

    It should be so obvious, why hasn't anyone had that idea yet, I have to wonder...

  • Is still a winning formula for some companies. Literally every company in the world that comes out ahead in this equation will be replacing humans as soon as possible. All those chat icons on sites? Having a real human on the other end will soon become a quaint thing of the past.

    Call center chat support and sales personnel will be the biggest losers.

    • Perhaps also the biggest winners. I'm a sales-related role. I spend much of my day patiently answering many of the same questions for customers and prospective customers. If ChatGPT could answer those questions, I wouldn't be out of a job. That would let me concentrate on the harder things that ChatGPT can't do. Now some of our lower-level sales people might find themselves replaced. But I would be able to close more business and be worth more money. We try to put this stuff into white papers and suc
      • That's sort of the point. Instead of needing 5 people I can get by with 2 people instead. Pretty sure most jobs that are just filling out forms could be done this way. So most banking, loans, insurance, call centers, simple tax filing (w2 only), and more. You'll still need people involved, but not nearly as many.

  • Are companies doing this or are people doing it?

    I'm a person who happens to work for a company and I tried using ChatGPT last week to write some code for a task that wasn't important enough for me to have automated yet. I found that the code ChatGPT provided wasn't correct, but it was pretty close. Even more interesting was that when I pointed out the errors, ChatGPT responded with "You're right, here's the correct code" and gave me something that worked.

    But on a different problem I found that it gave me ou

  • As far as I know chatGPT only works in the OpenAI playground and there is no API or pricing model for business use.

"The vast majority of successful major crimes against property are perpetrated by individuals abusing positions of trust." -- Lawrence Dalzell

Working...