Generative AI expected to replace 2+ million US jobs by 2030, those with higher education...

[midian182]

Forward-looking: Another study on the effect generative AI will have on jobs has been carried out, and it's a familiar story for white-collar workers: the technology will replace 2.4 million jobs in the US by 2030 and influence 11 million more. The (slightly) good news is that those job losses represent a lower figure than what many others have predicted.

Analyst firm Forrester's 2023 Generative AI Jobs Impact Forecast offers a similar conclusion to IBM's recent study: that while generative AI is going to cost many people their jobs, it will augment more roles than replace them. Forrester notes that 4.5 times more jobs will be "influenced" by the tech, which it defines as reshaping, retraining, and upskilling existing workers to incorporate generative AI tools into the daily workflow.

In the part of the report titled 'Let's Be Clear: Generative AI is Coming After White Collar Jobs,' it's stated that automation and AI overall will replace 4.9% of US jobs by 2030, with generative AI making up 30% (2.4 million) of those lost jobs. The likes of ChatGPT are predicted to replace 90,000 jobs in 2023, representing 9.3% of all jobs lost to automation and AI this year.

As we've heard before, the jobs most at risk of being impacted by generative AI include technical writers, social science researchers, research assistants, proofreaders, and copywriters. Harder-to-automate jobs with high generative AI influence, such as editors, writers, authors and poets, lyricists, and creative writers, are more likely to influence how jobs are conducted (via augmentation) rather than replace them.

Another finding is that the more education a worker has, the higher the influence generative AI will have on their job. Occupations with lower educational requirements like transportation and warehousing, construction, agriculture, and manufacturing will see less job influence, while those requiring bachelor's degrees or higher are going to see the biggest disruption.

It's a similar trend with wages: workers who earn under $60,000 are going to be less impacted than those earning $90,000 or more.

Job losses from generative AI are expected to be modest over the next couple of years as issues surrounding the tools are addressed, such as intellectual property rights, copyright claims, and plagiarism.

Forrester warns of potential AI roadblocks such as the prospect of generative AI producing "coherent nonsense" and the massive backlash against these systems from the public, as was seen with the Secret Invasion TV show's opening credits and the D&D book illustrations.

One of the earliest studies claimed generative AI could affect 300 million jobs globally. Forrester's conclusion is more optimistic, but it does warn that some workers will lose jobs they need, creating deep social challenges like those faced in the postindustrial Rust Belt. It's no wonder many tech execs are turning to drink and drugs to cope with the stress.

Permalink to story.

https://www.techspot.com/news/100085-generative-ai-expected-replace-24-million-us-jobs.html

 

[Uncle Al]

As previously predicted, the value of higher education is under attack, which promotes lesser and lesser education, making us vulnerable if and when AI fails and we have to rely upon educated people. This can be particularly dangerous as AI gets into the medical and legal fields. AI can be a very good thing but it needs to be used as an add-on rather than a primary tool. Now is the time when professionals may need to give serious consideration to unionizing in order to protect their jobs and futures. Hard to believe it has come to that .....

 

[Aranarth]

As previously predicted, the value of higher education is under attack, which promotes lesser and lesser education, making us vulnerable if and when AI fails and we have to rely upon educated people. This can be particularly dangerous as AI gets into the medical and legal fields. AI can be a very good thing but it needs to be used as an add-on rather than a primary tool. Now is the time when professionals may need to give serious consideration to unionizing in order to protect their jobs and futures. Hard to believe it has come to that .....

I'd love to see an AI fix a computer..

If the hard drive needs to be replaced or a computer wiped and reloaded you're still gonna need someone around who knows how to do it.

 

[Dimitriid]

I'm pressing X for doubt on this one: Don't get me wrong is not entirely wrong in the sense of the next 2-3 years of companies trying to cut some of those jobs but what isn't being told is how the next 3-4 years after that so many companies will experience catastrophic levels of failure and poor results due to AI incompetence to the point that it will crash the 'AI' bubble when it just cannot deliver on the results.

Think of it like a bigger, slower moving version of the (many) crypto/NFT craze: a lot of big companies like Microsoft and Nvidia really want you to believe in the AI future mostly because it is making them money by the truckloads but even right now today you can already find endless articles showing just how much AI flat out lies and how many problems it creates when it lies confidently.

And that's just the cases where the press can be a subject matter expert to fact check AI but if you fire your coders then AI starts coding itself into either unexpected failures, monstrously inefficient performance numbers or the most likely scenario, a massive avalanche of security issues well who's gonna fact check AI written code since you fired the coders? You're in deep trouble at that point.

It's not very exciting (read: good for the stock numbers of the companies heavily betting on AI) to temper expectations about the capabilities and limitation of Machine Learning models today by well, calling them 'AI' for once but outside of some creative areas (Because mainstream creative products haven't been very creative at all for years they cram it through with marketing) the world won't look all that different in 2030 in terms of AI, at most it will look different because of an AI created market crash that might severely weaken several major economies altogether if too many people trust the big player promises they cannot possibly keep.

 

[p51d007]

LOL, who cares...by the time THAT happens, some nut job will pop a nuke or EMP and wipe out anything with a chip and anyone that survives will be back to cave man days.

 

[Theinsanegamer]

New technology will be used to undermine high earning workers to the benefit of the 1%: example # infinity+1

 

[adamada72]

I believe this is the beginning of the end for a lot of jobs with very few new jobs to replace them.

 

[wiyosaya]

The sky is falling! The sky is falling! The sky is falling! 🤣

 

[Gastec]

The sky is falling! The sky is falling! The sky is falling! 🤣

Let the sky fall
When it crumbles
We will stand tall
Face it all together!

 

[kinetix]

The worst consequence that I see is what will happen is that almost everything will be nothing more than an interpolation and mixture, over and over again, of already existing knowledge. The new knowledge generated will be very scarce. If the higher educational levels are also the most affected, there will be fewer people in those categories (less incentives to improve), the educational level will be lower. And that cut/jump that would happen from "some college curses" directly to post-Doc makes me laugh...
and there are several other possible consequences.
Damn, Black Mirror is falling short, and I see once again how "The Space Merchants" is one of the books that has best predicted the future

 

[toooooot]

LOL, who cares...by the time THAT happens, some nut job will pop a nuke or EMP and wipe out anything with a chip and anyone that survives will be back to cave man days.

Cave men with survival skills and ranged weapons

 

[human7]

It's a bit too early to tell, and I'm not sure about some of the forecasts they made. For example, they put poets, editors, and creative writers on the low end of the automation potential, which is a severe underestimate (like it or not, LLMs are getting better and better at that work). They put technical writers on the opposite end, and coders in the middle. But we know that even things like Github copilot get things wrong a huge chunk of the time.

There's a lot of work that has gotta go into improving LLMs in order for some of their predictions to pan out, and I'm not sure they understand what can and cannot be automated within the next few years. Perhaps nobody does.

In theory, eventually anything we humans do can be automated (there's no physical law that says we cannot just invent an artificial human at some point in our future). Even so, it's a bit too early to call this doom and gloom forecast likely, especially since humans can improve themselves with the very technology said to be able to surpass them. Deus Ex style augmentations come to mind.

 

[Jack77]

The interesting part is that previous evolutions impacted blue collar workers, who politically have less influence. Now, white collar workers are at risk, the voting public of many parties in multi-party democracies. Will parties lose votes, will they react, or just keep following the lobbyists and how will that affect them?

 

[Lounds]

AI/ML should be a tool, not a replacement for trained professionals.

 

[wiyosaya]

The interesting part is that previous evolutions impacted blue collar workers, who politically have less influence. Now, white collar workers are at risk, the voting public of many parties in multi-party democracies. Will parties lose votes, will they react, or just keep following the lobbyists and how will that affect them?

When most everything that AI spits out has to be verified, I highly doubt these predictions will come to fruition. If they do, I bet that those contributing to this will realize that AI is nothing more than an attempt at a money grab by AI manufacturers.

Not to mention the immense toll AI is taking in energy and water use.

In the past, there were dire predictions about blue collar jobs due to robotics. Those predictions never came to pass and I doubt these predictions about AI use will come to pass either.

Let the sky fall
When it crumbles
We will stand tall
Face it all together!

Perhaps I should have added "/s" to my post?

 

[PEnnn]

I'd love to see an AI fix a computer..

If the hard drive needs to be replaced or a computer wiped and reloaded you're still gonna need someone around who knows how to do it.

And this "someone" is educated / studied computer science or similar.

The non- educated listening to their politicians (who publicly admit / promise they want to eliminate Dept of Education in the US) will lose no matter what. And deservedly so.

 

[Loadedaxe]

This is not the beginning. Technology has been taking over jobs for over a decade.

 

[Jack77]

In the past, there were dire predictions about blue collar jobs due to robotics. Those predictions never came to pass...

Technology has made it possible to cut out the more skilled blue collar workers, and to move industrial production to lower-labour-cost countries. That's a significant shift.

Generative AI is already cutting out technical engineers. And more jobs will follow. Bean counters do not care about correctness until it affects the bottom line. And then the impulse will be to try to fix things with a limited number of controllers and more technology. Resulting in a workforce of high-ups and as few as possible 'bodies', with as little as possible in between.

A lot of companies used to have some societal responsability and society building aspect to them. These are aspects I do not see in current corporations. And that will have consequences.

 

[Bluescreendeath]

As previously predicted, the value of higher education is under attack, which promotes lesser and lesser education, making us vulnerable if and when AI fails and we have to rely upon educated people. This can be particularly dangerous as AI gets into the medical and legal fields. AI can be a very good thing but it needs to be used as an add-on rather than a primary tool. Now is the time when professionals may need to give serious consideration to unionizing in order to protect their jobs and futures. Hard to believe it has come to that .....

As with a lot of current corporate cost vs benefit analysis, it might be cheaper to replace educated people in the medical and legal fields and then just pay out lawsuits from the medical and legal malpractice situations that happen.

 

[Robbie K]

Love how there were all these people saying it won't impact jobs.

 

[Psycho Killer]

Seems like this is an attempt to further erode the middle class, leaving the elites and the uneducated labor force.

 

[Bluescreendeath]

Seems like this is an attempt to further erode the middle class, leaving the elites and the uneducated labor force.

There is also the non elite portion of the upper class that makes up the majority of the upper class. Some studies showed that apparently the decline in the middle class was caused by a larger portion of the middle class joining the upper class than the middle class falling into the lower class.

Overall it still sucks that the middle class shrunk, but it wasn't all bad news.

 

[Fimbulvetr]

As a developer in large-scale ERP modules, today I tried using ChatGPT to write a program. I gave it very clear, concise instructions on what the program was to do, and it wasn't complex. I was rather shocked at what it produced - not at how good it was, but rather at how bad. The code it produced was so wrong, it had added non-existent fields into the code, ones that did not even exist in the underlying database tables, and then tried to use these fields in the code. It was utterly wrong, totally useless, and completely blew up when trying to compile/activate. What's more, the entire premise of the code was flawed - how it was going about trying to achieve the desired results would never have worked, which meant one could not even use the code as a skeleton or framework to assist in building a solution. Again, I was quite surprised. Maybe they'll get better, but judging from what it wrote today, it isn't a threat at the moment, at least not for what I do.