AI's grand promise: Less drudgery, more complexity, same (or lower) pay A report on occupational health warns that AI adoption may paradoxically increase workplace burdens rather than reduce them. As AI automates routine tasks, workers will shoulder new responsibilities: overseeing AI systems, catching their errors, and managing the resulting complexity – potentially triggering mental health …
So, instead of doing drudge work, the same amount of workers will spend the same or more time reviewing AI-generated drudge work.
But... if that's the outcome, why the hell would an organization deploy AI in those roles? What's the ROI, where's the benefit?
Besides satisfying investors' AI-FOMO, I mean. The capitalism algorithm is meant to work with rational actors...
That concept is sooo last millenium.
These days, it's all about paying insane costs for "CloudTM" and bragging about "AI integration".
Don't worry. The Board will soon realize that they can't give themselves their usual bonuses because their "AI"-integrated business is not making enough money for that anymore, so they will adapt.
How many workers will be laid off in the meantime is another issue . . .
Lol, allow us to introduce you to: Middle Management. Among other things, they have Budgets. If they don't use their full budget this year, it may be reduced next year. Spend!!!
OTOH, if everyone is becoming a team manager, then everyone will be payed a manager's salary! right? Right?...
On the serious side, though, the article proposes a world in which skilled employees are reviewing the un-skilled output of "Agents". It leaves out completely the process by which one _becomes_ skilled, and able to perform this review. No, no no, you can't *do* the work to learn it. Your job is to *know* the work and _fix_ the Actually Garbage output. If you don't know it already, we don't need to hire you. -> In about ten years, there is going to be another drop in certain skillsets (like COBOL - of the AI age) which will suddenly become high-paying (anything that was "agent-ized", and in the mean time everyone is going to go to smaller start-ups (now) and learn and do different things, as always happens.
You should be looking to start-ups and smaller companies. That's where your skills will develop, and where you'll find jobs, and teams and people who need you and appreciate your presence. (They're people though, so certainly don't approach as though you're their savior - everything is ok, now that they have You.)
The biggest risk of Agentization is: it can't do the job, perfectly. You're stuck with the agents, they're part of the process. You're stuck with the skilled labor: you need people to check it and correct it. On top of that, you're *also* stuck training and developing all the people you had to train and develop before - so that you can ensure the agentic output continues to be correct and get fixed. Agentization is double-cost. Ouch.
The biggest risk of Agentization is: it can't do the job, perfectly. You're stuck with the agents, they're part of the process. You're stuck with the skilled labor: you need people to check it and correct it. On top of that, you're *also* stuck training and developing all the people you had to train and develop before - so that you can ensure the agentic output continues to be correct and get fixed. Agentization is double-cost. Ouch.
This is where the smart beancounter might spot an opportunity for cost cutting. But will they cut the right cost?
In my experience, bean counters are often quite bright. Their problem is that they usually have to operate in the quite bizarre landscape between marketing (lying scum), management (politicized and too often rather demented), regulations (which assume a sane framework) and reality (inconvenient and inflexible).
Why would organisations deploy it? The answer is the premise of the article, lower cost of compensation. Workers leaving would logically have reduced loss of knowledge too.
Assuming the article holds true, AI will eliminate more than half the back office roles in the public sector for example, allowing taxes to fall and so productivity to rise. Coupled with agentic so (stop laughing, it's not my premise) being able to run unattended for a time, strikes become a thing of the past too, this further increasing output.
Of course, the article could have all this very wrong indeed. I'm pretty sure ai will radically alter work. It's the foreseeing how that's the difficulty today.
"This effectively transforms workers into managers of AI systems – a role not everyone is suited for."
If the rationale for adopting Dunning-Kruger machines is to lay off the workers then it will be the managers who are now the managers of AI systems, also a role they're probably not suited for.
how sad.
in the process of swapping from doing things to teaching ai doing things for you, ai will be more fluent in doing what you are paid to do, up to a point when you are no longer paid.
as sharpening the knife that will be stabbed into you so you could die more quickly.
agentic ai is still in early stage, but with the collective wisdom of poor fellow workers it will definitely become very capable, and ultimately more wealth and control for sam altman et al.
Well, that sort of depends...
I'm very good a training my replacement in a redundancy situation and I put a lot of effort into it too. But I'm not a miracle worker and those pulling the trigger usually end up reaping the whirlwind.
Given enough time and a good enough AI, I'm sure I could train it up. But will the AI be good enough? No. Will I be given enough time? Also no. It is what it is.
Which part of the inviolable law of economics regarding monkeys and peanuts did these genius MBAs not understand ?
One imagines these fools would opt for the cheaper nearly qualified brain surgeon to excise their brainstem tumour — sadly I really think they would.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
