Jobs Vulnerable to AI Replacement Actually 'Thriving, Not Dying Out', Report Suggests (fortune.com) 42
AI startups now outnumber all publicly traded U.S. companies, according to a year-end note to investors from economists at Vanguard.
And yet that report also suggest the jobs most susceptible to replacement by AI "are actually thriving, not dying out," writes Forbes: "The approximately 100 occupations most exposed to AI automation are actually outperforming the rest of the labor market in terms of job growth and real wage increases," the Vanguard report revealed. "This suggests that current AI systems are generally enhancing worker productivity and shifting workers' tasks toward higher-value activities..."
The job growth rate of occupations with high AI exposure — including office clerks, HR assistants, and data scientists — increased from 1% in pre-COVID-19 years (2015 through 2019) to 1.7% in 2023 and beyond, according to Vanguard's research. Meanwhile, the growth rate of all other jobs declined from 1.1% to 0.8% over the same period. Workers in AI-prone roles are getting pay bumps, too; the wage growth of jobs with high AI exposure shot up from 0.1% pre-COVID to 3.8% post-pandemic (and post-ChatGPT). For all other jobs, compensation only marginally increased from 0.5% to 0.7%... As technology improves production and reallocates employee time to higher-value tasks, a smaller workforce is needed to deliver services. It's a process that has "distinct labor market implications," Vanguard writes, just like the many tech revolutions that predate AI...
"Entry-level employment challenges reflect the disproportionate burden that a labor market with a low hiring rate can have on younger workers," the Vanguard note said. "This dynamic is observed across all occupations, even those largely unaffected by AI..." While many people see these labor disruptions and point their fingers at AI, experts told Fortune these layoffs could stem from a whole host of issues: navigating economic uncertainty, resolving pandemic-era overhiring, and bracing for tariffs. Vanguard isn't convinced that an AI is the reason for Gen Z's career obstacles.
"While statistics abound about large language models beating humans in computer programming and other aptitude tests, these models still struggle with real-world scenarios that require nuanced decision-making," the Vanguard report continued. "Significant progress is needed before we see wider and measurable disruption in labor markets."
And yet that report also suggest the jobs most susceptible to replacement by AI "are actually thriving, not dying out," writes Forbes: "The approximately 100 occupations most exposed to AI automation are actually outperforming the rest of the labor market in terms of job growth and real wage increases," the Vanguard report revealed. "This suggests that current AI systems are generally enhancing worker productivity and shifting workers' tasks toward higher-value activities..."
The job growth rate of occupations with high AI exposure — including office clerks, HR assistants, and data scientists — increased from 1% in pre-COVID-19 years (2015 through 2019) to 1.7% in 2023 and beyond, according to Vanguard's research. Meanwhile, the growth rate of all other jobs declined from 1.1% to 0.8% over the same period. Workers in AI-prone roles are getting pay bumps, too; the wage growth of jobs with high AI exposure shot up from 0.1% pre-COVID to 3.8% post-pandemic (and post-ChatGPT). For all other jobs, compensation only marginally increased from 0.5% to 0.7%... As technology improves production and reallocates employee time to higher-value tasks, a smaller workforce is needed to deliver services. It's a process that has "distinct labor market implications," Vanguard writes, just like the many tech revolutions that predate AI...
"Entry-level employment challenges reflect the disproportionate burden that a labor market with a low hiring rate can have on younger workers," the Vanguard note said. "This dynamic is observed across all occupations, even those largely unaffected by AI..." While many people see these labor disruptions and point their fingers at AI, experts told Fortune these layoffs could stem from a whole host of issues: navigating economic uncertainty, resolving pandemic-era overhiring, and bracing for tariffs. Vanguard isn't convinced that an AI is the reason for Gen Z's career obstacles.
"While statistics abound about large language models beating humans in computer programming and other aptitude tests, these models still struggle with real-world scenarios that require nuanced decision-making," the Vanguard report continued. "Significant progress is needed before we see wider and measurable disruption in labor markets."
Fishy (Score:5, Insightful)
This piece sounds like whitewashing to me. "Look at all the good AI is doing for us". "AI is actually our friend."
Re:Fishy (Score:4, Insightful)
Re: Fishy (Score:2)
You shouldn't talk to the goth kids. They'll try to make you one of them.
https://i.redd.it/w577feemjtag... [i.redd.it]
Re: (Score:2)
Re: (Score:2)
Think about it. That's GREAT for the economy, because it would increase demand for workers and work.
It's analogous to software bloat. Remember "Nobody will ever need more than 640k of memory?"
How about the paperless office?
Growing the economy is a Religion. Chant the following: "The economy MUST grow".
Things like global warming and pollution are irrelevant, because we MUST GROW the economy.
From that point of view "AI" effing
Re: (Score:2)
That argument is only 50% of the cost calculation. And for some fields (software) we already have numbers saying that not only does it make people slower, the results are worse. One problem is that LLMs make a lot of security bugs and chances are that more will slip though that code by a competent person did contain in the first place. Another problem is that LLMs cannot do software architecture. And another is that beyond a pretty low complexity level, they get confused.
For the lawyer, things are open at t
Re: (Score:2)
I don't see AI as a tool, because it is inconsistent and unreliable. You get a different answer every time.
A hammer's functionality doesn't change on a minute-per-minute basis, and neither does the functionality of a forklift, a lathe, or a CNC machine. Even computers are generally expected to have consistent, reproducible behavior, and we get mad when they don't. Those are real tools.
A long time ago, I was a fork truck driver. When we had trucks that were based solely on hydrolic valves, they were terr
Re: (Score:1)
There is so much propaganda that a thoughtless AI can rephrase it all into infinite versions of "smoking is good for you" on top of actual marketing people who will promote and lobby with more resources than anything except a presidential campaign. They were involved in that too.
They only preached "regulate us" when it was a good hype campaign to raise investment and become 1st to monopolize the market. Pretty sneaky being a non-profit tech startup and brilliant on the fake selfless act. Altman is a next le
Re: (Score:2)
Yep, and those aren't high-vulnerability jobs, making this a strawman argument - some example of high-vulnerability jobs would be graphic artists, voice actors, and translators.
Re: (Score:2)
Having experienced some AI "translations" and "voice acting" and "graphics", I think they are not at risk to any significant degree in the longer term. Replacing competent work with slop takes a while to be reflected in revenue, but it will be.
Re: (Score:2)
AI is neither your friend nor your enemy. It is a tool and an amplifier. People with skill can do twice the work with it, people without can create twice the chaos with it.
Re: (Score:2)
The question is whether it is actually a beneficial amplifier. It does amplify errors, bad style, genericity, and outright fabrications as well. That may kill the use or even make the usefulness negative. One example: https://mikelovesrobots.substa... [substack.com]
It is quite possible other areas will be affected similarly. I mean, yes, you can probably replace an incompetent and unreliable employee with LLM-type AI, but why were you employing that person in the first place?
Re: (Score:2)
I had the discussion here some time ago. The Shovelware article is obviously missing some other factors.
The line of argument sounds good, but the premise is that AI cannot create this kind of cheap games, while everyone with a mid-range GPU can produce such games with freely available models. So there must be another factor preventing shovelware ... or people overlooking the shovelware that exists and may not be promoted by the store algorithms.
Using Qwen Coder I can produce a HTML5 Tetris in a single promp
Re: (Score:2)
With reasonable skills it will save you time.
That seems rather questionable. Or something you actually have to prove, not just assume.
Re: (Score:2)
You can test it yourself. I assume you have the skills to rate the outputs and one can learn how to steer the things to what you want rather quickly *if* you know what you want. The problem are the people who cannot judge if the output is what they want and think they can go with the flow without giving own directions.
Re: (Score:2)
"You can prove it yourself" is a deflection, not proof. Do better.
Re: (Score:2)
We both know, that you are very skeptical when it comes to AI and I think you would doubt reports of others, so the only way is to try it yourself.
Re: (Score:2)
To me it more sounds like "the numbers do not pan out" and "AI is not performing on the level people think it performs". AI replacing workers is only a bad thing if it actually works to a significant degree for that. It does not seem like it does that. There are other bad aspects to LLM-type AI, no argument. But this time, the grand promises will (again) probably not pan out.
A little early to declare trends (Score:5, Insightful)
ChatGPT was released in late 2022, just over 3 years ago. It takes a little longer than 3 years for a new technology's impact to be clearly seen. As with any trend, there are ups and downs all the time, it takes time to establish a "trend."
Re: (Score:1)
Because AI isn't working out (Score:1)
Those jobs continue to thrive because AI isn't reliable and gives incorrect results.
They are still working on the build out (Score:1, Troll)
Automation devastated the middle class in America from 1980 on. It was factory automation. It ripped through high paying factory jobs and that trickled through the rest of the economy as those high-paying jobs went away and were replaced by lower paying lower skilled mcjobs.
We just don't talk about it. Everyone points to the blasted out hellscape
Re: (Score:2)
People never see it coming and most have an optimist bias causing them to stark escape realities until it is unavoidable and at that point they're too ignorant, scared, and shocked to think properly. Ripe for propaganda; today, it's the best most evolved propaganda for sale along with the best distribution systems for a public with the shortest of attention spans. Attacking the messenger is the default of most people for the whole process.
Nobody seems to have learned from Big Tobacco and how incredibly inf
Re: (Score:2)
Attacking the messenger is the default of most people for the whole process.
Yes. Abject stupidity all around.
Re: (Score:2)
If you're under 65 you're going to see 25% unemployment in your lifetime.
It's looking like we'll have world war well before that. Who can say how that will shake out?
Re: (Score:2)
This is not comparable. Automation generally removes some flexibility (more recently, it may actually do the opposite), but automation does not remove reliability. In fact, it generally improves it by more standardized products. LLM-type AI kills reliability. And that is not really fixable as it is baked into the math.
For special purposes, it will probably be possible to create small dedicated LLMs and fit them with a result-checker (that uses deduction and hence is reliable), but these will look very diffe
Seems at odds with reality (Score:2)
There's been a lot of layoffs. Where are we seeing actual job growth? It's hard to detect since the press often ignores hiring sprees but still reports layoffs.
Re: (Score:2)
I don't have mod points to mod you down, so I'll respond to your stupidity about "the press". The press reports whatever government report comes out about jobs and employment. Every month the press reports this.
Here is a report [cbsnews.com] from the press two weeks about employers adding jobs. There aren't any numbers to report for Ocober because Orange Man had the government shut down so data couldn't be collected. In about two weeks the press will report on the December jobs numbers.
If you can't be bothered to read
Re: (Score:2)
First off, check the needless hostility.
Secondly, I'm not talking government jobs data, but rather stories like these [techcrunch.com]. Do yourself a favor and try reading next time.
Inside the above link you'll find a breakdown of which companies were laying off people and how many. Do press outlets show the same hiring data?
No, they don't.
Re: (Score:2)
I see at least two reasons for the press to report layoffs and not hiring.
1. "If it bleeds, it leads." Long before social media, newspapers focused on articles that engaged and got your attention. And we as humans tend to be more engaged with stories of extremes and unusual situations, especially tragedies. People are naturally risk-averse. Layoffs fit the narrative of what interests us to read.
2. The easiest way to win a race is if nobody knows you're racing them. Businesses don't announce big hirin
Re: (Score:2)
Where are we seeing actual job growth?
It is right there in the title of the story. Seriously.
At a lower cost. (Score:3)
Bank Teller was a respectable job until the ATM came. The ATM ushered in more but worse banks and tellers. I assume ( and observe that ) it's the same with software development.
Re: (Score:2)
From the known numbers, software development gets slower (!) with LLM "help". At the same time, software now often is so abysmally bad that it becomes unsustainable to use it, and that already without LLM "help". The ones that actually benefit from LLM help when creating software are the attackers as they do not need reliability, security or maintainability for attack code.
Hence I predict that in the software-space, a really big shakeup and probably liability and regulation are not very far ahead.
Quantity Over Quality (Score:1)
Re: (Score:2)
The feds should just issue everyone their own corporation. FU-corp?
Fair to State that Slashdotters Know Better (Score:3)
bullcrap (Score:1)
AI Generated Report States AI Does No Harm (Score:2)