Forgot your password?
Close
typodupeerror
AI

AI Is Hitting Entry-Level Jobs Hardest, Stanford Study Finds (arstechnica.com) 64

Posted by BeauHD from the would-you-look-at-that dept.
An anonymous reader quotes a report from Ars Technica: For years, AI industry watchers of all stripes have been warning of a coming jobs apocalypse driven by ultra-intelligent AI systems that will be able to replicate most human tasks more cheaply. Now, newly updated research from Stanford University economists suggests AI seems to be causing significant entry-level job losses for younger workers in some fields, even as older workers appear largely unaffected so far.

The August 2026 edition of "Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence" updates and revises a paper of the same name published last year with fresh data and refined statistics. In that update, the Stanford researchers find the employment trends they identified for entry-level workers last year are persisting and expanding. Specifically, employment levels for workers ages 22 to 25 in the most "AI-exposed" occupations are now 19 percent below those of their peers in fields less exposed to AI disruption. Last year, that gap measured just 13 percent.

[...] Digging deeper into the data, the researchers found that this phenomenon is mainly manifesting itself through lower hiring rates for entry-level workers in AI-impacted fields, rather than increased firings or employees quitting. They also found that the labor market effects among this age group were mostly seen in lower overall employment, rather than reduced pay rates. But not all jobs that show potential for AI "disruption" are created equal, the researchers found. In its Economic Index, Anthropic differentiates between queries related to tasks that are "automative" (i.e., fully replacing work previously done by a human) or "augmentative" (i.e., helping human workers be more effective at tasks they are still needed for). By this measure, jobs like "accountants and auditors" and "receptionists and information clerks" were among those judged most susceptible to AI automation, while jobs like "chief executive" and "registered nurse" were among those using AI augmentation most often.

Unsurprisingly, jobs where AI automation is prevalent are the ones showing the worst relative employment levels for entry-level workers these days. "The findings are consistent with automation-oriented uses of AI substituting for labor while complementary uses are associated with flat or rising employment," the researchers write. Interestingly, the researchers found that the impact is strongest in entry-level jobs built around "codified" knowledge, which is the formal, documented skills that AI can more easily replicate. Meanwhile, experienced workers in roles relying on tacit, practice-based knowledge have seen stronger employment growth.
This discussion has been archived. No new comments can be posted.

AI Is Hitting Entry-Level Jobs Hardest, Stanford Study Finds More

AI Is Hitting Entry-Level Jobs Hardest, Stanford Study Finds

Comments Filter:

  • cutting entry-level is bad long term

    • Re: (Score:2, Troll)

      by rsilvergun ( 571051 )
      We stopped caring about long term in the 80s. All that matters now is who's the next trillionaire and protecting the sanctity of women's sports

    • Re: (Score:2, Funny)

      by HuskyDog ( 143220 )

      Well, presuming that you mean from your refreshingly terse comment that if you don't recruit entry level roles then in the longer term you will not have any experienced staff because obviously they were all entry level staff once upon a time, then I'm not sure that I agree because AI will presumably get better over time.

      If you don't recruit any entry level staff then obviously in five years time you won't have any staff with five years experience. But, perhaps in five years time (or possibly less) AI will

    • cutting entry-level is bad long term

      Hiring Didn't Earn It, was bad short term. And middle term. And long term.

      As if we didn't notice, what ruined entry-level for the rest. Time will tell if we adjust HR hiring practices well enough to bring the FNG back.

    • cutting entry-level is bad long term

      Entry level is undermining itself. The vibe code, which results in two things. (1) They don't learn much. (2) More senior devs have to intensively review everything, Its not really saving them any time. They might as well just apply a test driven development methodology and have documentation and prompts defining constraints and the testing. In short the seniors applying their experience to provide direction to the AI assistants rather than apply their experience to reviewing the schlop the vibe coding juni

      • They are being ordered to use AI to increase their productivity by management which can't do their job and therefore doesn't understand that this will also increase their error rate.

        • Re: (Score:2)

          by drnb ( 2434720 )
          Actually the last one I observed was not ordered to do anything. However he seemed to believe that his internship was a personal quest to test the limits of vibe coding. We are reconsidering our long tradition of offering internships to the local university. The next intern will be warned about the mess his predecessor created. We'll see how the next one goes.

  • Entropy phase (Score:5, Interesting)

    by geek ( 5680 ) on Tuesday August 25, 2026 @02:12PM (#66306474)

    Ancient Rome made amazing aqueducts. One day they had built all the ones they really needed and with no new projects for a new generation to learn from, they quickly just forgot how to make aqueducts.

    "Big Tech" is now entering this phase. In a generation all innovation will be gone.

    • Re:Entropy phase (Score:5, Insightful)

      by drinkypoo ( 153816 ) <drink@hyperlogos.org> on Tuesday August 25, 2026 @02:32PM (#66306510) Homepage Journal

      The problematic thing about that analogy is that a lot of the aqueducts are still standing, and some are even still in use, but the things being made by big tech today won't still be standing in a couple of decades — let alone a handful of centuries. Even if you lop off a couple of orders of magnitude to account for the rate at which technology accelerates it's still looking bad by comparison. What the tech industry is doing right now is more closely akin to tearing down the aqueducts than to building any.

      • The problematic thing about that analogy is that a lot of the aqueducts are still standing, and some are even still in use, but the things being made by big tech today won't still be standing in a couple of decades — let alone a handful of centuries. Even if you lop off a couple of orders of magnitude to account for the rate at which technology accelerates it's still looking bad by comparison. What the tech industry is doing right now is more closely akin to tearing down the aqueducts than to building any.

        Not quite adequate. More along the lines of tearing down the aqueducts, creating technological barriers to prevent knowledge of how to build new aqueducts from being accessed by anyone who hasn't submitted fealty to them, then going out of their way to destroy the previously available knowledge that led to the first aqueducts so that "the wrong people" don't try to create any without the tech broligarch blessings. And, with an extra bit of kick in it, working every possible angle they can to promote destroy

    • If what you were suggesting were true we'd all be wearing sackcloth and bemoaning the complete lack of sanitation. If the big tech companies stop innovating they'll be replaced by something else that does, just like human civilizations largely managed to advance even though Rome declined and eventually fell.

      I have no loyalty to the current tech companies beyond their ability to produce products or services I want. The industry is already full of corpses from past decades that failed to innovate. Given lo

    • Ancient Rome made amazing aqueducts. One day they had built all the ones they really needed and with no new projects for a new generation to learn from, they quickly just forgot how to make aqueducts.

      "Big Tech" is now entering this phase. In a generation all innovation will be gone.

      Already starting to see it with the Cobol codebase. Hardly anyone goes into that area because it's seen as a dead end. Yet vast portions of the economy still run on Cobol, it's very expensive to replace it, all the grey beards that can maintain it are retiring, and there is a no replacement pool for those folks. I still know folks that miss the days when if you needed a custom accounting process or report change, you just took the lead programmer a batch of fresh baked cookies, cause that got you to the

      • Re: (Score:2)

        by drnb ( 2434720 )

        Already starting to see it with the Cobol codebase.

        A skilled software developer can learn Cobol in months. The requirements discovery, problems solving, and design skills are the difficult parts of software development, not implementation details like the programming language.

    • Ancient Rome made amazing aqueducts. One day they had built all the ones they really needed and with no new projects for a new generation to learn from, they quickly just forgot how to make aqueducts.

      "Big Tech" is now entering this phase. In a generation all innovation will be gone.

      Big Tech was relevant when there was actually Tech to invent. Today, it all seems like same shit solution to a non-problem, different app wrapper.

      Because it is.

      Thank a patent war chest whore for strangling innovation to death with litigation. Plenty Too Big To Fuck You proved why it's no longer worth trying to invent anything anymore.

    • Just as we no longer need Roman-style aqueducts, in the tech future we will no longer need certain kinds of tech. For example, Q&A websites are largely irrelevant now. Why would we want to go back to such sites? Like the aqueduct builders, we no longer need Q&A website builders.

  • Stanford Economist Now Believes an AI Job Apocalypse Is Unlikely [slashdot.org].

    Economists, at this point doesn't everyone just laugh at them?
    • There is a nuance between an AI apocalypse and significant job losses hurting young people in the economy... And you too by the way because your social security and Medicare are going away.

    • Economists, at this point doesn't everyone just laugh at them?

      Is it the same article, just with a different headline (A/B testing clicks and headlines)?

      Regardless, we are likely in a recession and companies are not hiring because they can't afford it, and are claiming they can replace that work with AI.

    • Economists, at this point doesn't everyone just laugh at them?

      It's kind of like laughing at the weatherman when he's wrong. Most people aren't stupid enough for that to make them feel smart, but you do you.

  • A few days ago, when Slashdot reported this study https://it.slashdot.org/story/... [slashdot.org]

    The headline read Stanford Economist Now Believes an AI Job Apocalypse Is Unlikely

    So, one angle says not a problem. But the other angle says that it is a problem. Entry level gets hit, that means that there is no one to become mid to high level. Then what?

    • Re: (Score:2)

      by ceoyoyo ( 59147 )

      Mid level becomes entry level. Of course entry level jobs are going to be the ones replaced first. They always are.

    • A few days ago, when Slashdot reported this study https://it.slashdot.org/story/... [slashdot.org]

      The headline read Stanford Economist Now Believes an AI Job Apocalypse Is Unlikely

      So, one angle says not a problem. But the other angle says that it is a problem. Entry level gets hit, that means that there is no one to become mid to high level. Then what?

      The flaw in the assumption that suddenly people won't know anything is that entry level changes, technology develops. People will be hired who will develop as the technology changes. This is a process that has gone on for a long, long time.

  • The automative vs. augmentive distinction that AI companies and the ownership class want you to believe in is nonsense meant to keep the working class calm as their job opportunities dwindle away. The only difference between them is a quantity threshold. Augmentive means multiplying the number of jobs available by a number greater than zero but less than one (as fewer people can do the same work), automative means multiplying the number of jobs available by zero. Both mean jobs are lost, the only difference

  • The young one know how "job" works mostly from watching enactments on video,
    and expecting a TV work fantasy environment, getting to define their role to play there.
    This is where the slack doesn't pay off like they had thought.

    • Whose fault is that? They didn't raise themselves. Parents keep fucking up and then blaming the product of their fuckups.

      • Whose fault is that? They didn't raise themselves. Parents keep fucking up and then blaming the product of their fuckups.

        Yes, but not just parents. Schools, media, pop culture. All had a part. And the public schools had an outsized effect.

        I managed to keep my son away from the worst of it. Kept him off the drugs the schools wanted him on, taught him that self esteem comes from accomplishment, not because the teacher told him he was the center of the universe, and he had great potential. He's doing okay. His Ritalin drugged up friends? Not quite as well.

        And the worst thing we taught the kids is in your answer. It's alway

        • Conservative politicians made the schools shit and conservative voters supported that, now want to cry that the schools are shit. Cry harder.

          • Conservative politicians made the schools shit and conservative voters supported that, now want to cry that the schools are shit. Cry harder.

            Did conservative politicians decide that almost all young males needed to be put on Ritalin? Or that teh solution to self esteem issues was teaching children they needed to be rewarded for getting their shoes on the right feet? or that Or that male teachers needed phased out because all men are sexual predators? Or that schools can assist in transitioning children to a different "gender," and that the child's parents have no right to be told? Yeah, that was all conservatives causing all problems.

            Your hat

    • Re: (Score:2)

      by gweihir ( 88907 )

      Nonsense. That is just the ages old complaint of old, arrogant, and pretty incompetent (for their age) assholes about the youth. I do not seem my students working any less hard than I did when getting my degree. Yes, they have a lot less experience and yes, there are a lot of things I can do they cannot do or cannot do well. But give them the 35 years of experience I have on them and that will chance.

      Obviously, there are lazy, arrogant and plain dumb students and generally young people today. But the same

  • AI can replace the jobs that can be done by idiots?!?

    • Re: (Score:2)

      by gweihir ( 88907 )

      That is total nonsense. The idiot here is you. What LLMs can do, to some degree, is replace inexperienced smart (!) people, but only if driven by an experienced smart (!) person. Inexperienced smart people are however the only known source of experienced smart people, and you need to give them 10-20 years of work experience to get there.

  • Not surprised. (Score:3)

    by whoever57 ( 658626 ) on Tuesday August 25, 2026 @03:31PM (#66306598) Journal

    I am not surprised at this.

    In my use of AI, I realize that the user still needs to ask the right questions, with sufficient detail to get a good result. The user needs to push and probe on architecture to get a good result. This is not entry-level knowledge.

    AIs provide a force multiplier to people who already have the background information to ask the right questions. These are the people who would architect solutions and then pass off implementation and guidance to the entry-level people.

    • Agreed, my experience mirrors yours. I do think that bright young users could get good use of AI, just as I've worked with bright young college interns that could be a real help in coding without AI. Unfortunately, the majority are not "bright." That's not a slam, just how the normal curve works.

  • I'm going to approach this from an engineering angle, what does a Jr Engineer really do? They assist the Sr Engineers, and what Sr Engineers have Jr Engineers do? The grunt work, the mindless work, the stuff we're too busy to be bothered by, and that's both a shame, but understandable.

    I have a Jr Engineer under me right now, she's great! I give her an enormous amount of self-driven work, so she can explore, have fun, create, and improve products. She can play around, she can make a mess, but, creating a

    • Re: (Score:2)

      by gweihir ( 88907 )

      AI has removed that, and, but in the process in 20 years, when I retire, whose going to pick up that torch and run with it? When you hit a problem, AI can't solve, or doesn't solve correctly, how will you know? We need to keep hiring for Jr roles, Jr roles lead to Sr roles, if you don't have the Jr staff coming in, in 10 years, 15 years, 20 years, we're fucked!

      The companies that stop educating entry-level engineers are dead. Sure, they will only die in 10-20 years, but they now have a cancer that is slow-growing but incurable. There is always a shortage of competent senior engineers. That is the only reason the industry is (was) educating them with high effort. They are direly needed, cannot be replaced by anything else (no, LLMs will never be able to do it) and they are rare. You cannot get them from other places either. All you can do is give somebody with a go

      • Exactly, it might be more important than ever to have a healthy and strong entry-level base to pull from. Think about ChatGPT generating a kernel, what do you learn from that? Sure, you can study the code, learn what works, why and how, comment the living hell out of it, and really use it as a learning base, but, how many companies are going to fund that? You want two months to review generated code, that works, why? In 20-years if some of those boards are still operational, who are they going to call to

  • hmm (Score:2)

    by nomadic ( 141991 )

    I feel bad for entry-level folks -- though continually surprised how many of them outsource so much of their thinking to AI themselves. It seems like most college students cheat with AI, making them less useful as entry-level employees.

  • I'm a software developer but I've always had misgivings about university curricula for professional/technical tracks. Maybe if we looked at education again as teaching people to read challenging books, write well, and think critically, we would all be better prepared for whatever life threw at us.

    A few years before I enrolled at Vanderbilt, all freshmen had to take calculus. And even as an engineering major, I was expected to pass tough humanities classes.

    When you earned that diploma, prospective employers

  • No new workers, dumb framework clowns in the middle, and all the real devs aging out?

    Welcome to the Temu Skynet future!

  • I know, while I was still working a day job, that I saw a number of utterly stupid job ads for "entry leve", "must have 2 or 3 or 5 years experience".

  • In common terms, LLMs are not intelligent. They can, with some probability of gross error, do simpler basically generic tasks on verbal instruction. That is what entry level jobs are also doing to get people to higher competence levels. That works for people because people have insight and develop understanding. LLMs cannot do either of these. Replacing entry-level jobs with automation does not make any sense, since they are never cost-effective anyways. Productivity is not their purpose. They are designed

Slashdot Top Deals

Our business in life is not to succeed but to continue to fail in high spirits. -- Robert Louis Stevenson

Close