Betting on AI leads to more educated juniors, fewer mid-level bosses – study When US companies invest in skills associated with artificial intelligence, they tend to gain more educated junior employees while shedding middle-management and senior roles, according to newly published research. That is to say, AI increases advanced degrees among staff while flattening the organization. In a paper titled …
Junior employees have a higher level of educational attainment because today you can't get a career without a degree. Companies today require a degree for junior positions that in the past a good academic achievement at school (O Levels) or sixform (A levels) would have been enough. Because not that many people entering the job market for the first time had degrees. As there were nowhere near the number of degree places, universities or shear variety of degree subjects.
When I left school in the 1970s very few people went on to degrees, to get say a job in a drawing office (draftsman) or a career in retail management good O Levels or wait 2 year for good A Levels would have got you in. Then it became A levels and today it will be vocational degree in drafting or retail management.
AI or not, juniors are going have degrees because without them they stand no chance of getting the first job of a career.
As for the "with greater emphasis on STEM degrees", I wonder what qualifies as STEM are they including social sciences as they do in the US or are they excluding them as we do in the UK.
I agree that what you describe is correct, but the paper is looking specifically at companies that invest in AI, and is finding they have more young graduates than the others. The phenomenon you describe would apply equally, so there's some other factor at play.
Actually, I'm not surprised that companies investing in AI have more young graduates, because if you want to use AI you'll need computer scientists and they don't grow on trees. I'm more interested in the other bit, how they have fewer middle managers. I wonder whether perhaps the causation goes the other way around: maybe companies with lighter management are more likely to embrace AI.
@Filippo "I agree that what you describe is correct, but the paper is looking specifically at companies that invest in AI, and is finding they have more young graduates than the others."
From my read of the paper they are not comparing companies that invest in AI tools against companies that have not. They are looking at companies that have invested in AI tools and found that they have more graduates than they did before in earlier years. They then conclude that the increase in graduates is due to AI and unique to AI, that companies embracing other technologies over the years have not seen that effect.
But is that conclusion right or are they seeing what is happening with all companies. Companies today demand a higher educated workforce both for existing junior roles and new roles created, due to them embracing new technology/tools, AI is just one more new technology.
The paper excluded tech companies that create the AI systems, it only looked at companies that used the AI tools that tech companies created. Sure you need computer scientists to create these AI tools, but do you need computer scientists to use AI tools? Ultimately no, they are just tools, though today with these LLMs maybe, they are new*. But in time computer scientists will not be needed.
@Filippo "I wonder whether perhaps the causation goes the other way around: maybe companies with lighter management are more likely to embrace AI."
You could be right. ^ :)
* By new I mean new as a tool used by non tech companies.
If your company moves towards AI, you need staff who can do it. They are much more likely to be young, and to be taken on in a junior role, since the field is only 10 or so years old.
So adding young people to the payroll reduces the average age and treating them as junior staff reduces the proportion of senior managers. But maybe you need a STEM degree to appreciate that.
Nope. They were talking about mid-level managers, not senior, so their point stands: Model a filesystem with 8 records per inode, then do it again with 32 per inode, and compare the depth of each tree...
In the companies they surveyed who “used AI” (whatever that means, which is problem #1) they found that the ratio of junior staff to middle managers (i.e., the average size of each managed group) was higher than in other companies. This meant fewer middle managers were required for the same number of juniors than would have been in companies in the control survey (whether there was a proper control group would be problem #2), leading to a flatter organisational structure.
At my first glance, the problems with this survey are more to do with shaky methodology than spurious interpretation of the results.
"That is to say, AI increases advanced degrees among staff while flattening the organization"
That's because A,I as it is done today, is "new", in twenty years you will have aged staff again.
Also you should probably hire some people with degrees in philosophy and ethics just saying.
It may save you from billion dollar lawsuits in the future.
《hire some people with degrees in philosophy》
Actually pretty good advice even leaving questions of ethics and morality aside, which I don't, as this discipline considers the nature of knowledge, belief, truth etc etc as well as formal logic in its many forms. As an example, I very much doubt many STEM students are ever exposed to modal logics or any formal logic.
I would have thought AI / ML would be seeking firmer foundation on models of knowledge, belief etc and apply reasoning techniques from various logics. To my mind, a formal model and associated logic(s) of time would be essential for reliable, consistent reasoning. eg that cause precedes effect ie a model of causality.
OpenAI's explanation of its recycled joke "Why did the tomato turn red? It saw the salad dressing." speaks volumes.
Like a four year old's confabulated explanations after seeing "adult" behavour or other farm yard antics. The layers of socialization built on top of the foundations of human biology appear to be completely absent.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
