Many employees are using AI to create 'workslop,' Stanford study says Workers are getting lazy about using AI to do their jobs for them, and the results are both costly and increasing distrust in the workplace. An ongoing study by Stanford's Social Media Lab and behavioral research business BetterUp Labs says 40 percent of US workers are reporting AI-generated garbage, dubbed "workslop," coming …
Hades reserved this particularly maddening punishment for us due to our hubristic belief that our cleverness surpassed that of Zeus himself, skyrocketing into the realm of so-called SuperIntelligence ...
Condemned then we are to endlessly detangle the AI-generated workslop garbage that we've created for ourselves, with more of it produced each time we think the job finally done.
May the gods of computing have mercy on us as we battle this monstrous effluent that flows both ways, like a continuously exploding commode ... imho!
We created this Gordian Knot. Who will have the courage to wield a sword?
Morpheus: We have only bits and pieces of information, but what we know for certain is that some point in the early twenty-first century all of mankind was united in celebration. We marveled at our own magnificence...as we gave birth...to A.I.
Neo: A.I. - you mean Artificial Intelligence?
Morpheus: A singular consciousness that spawned an entire race of machines. We don't know who struck first - us, or them. But we know it was us that scorched the sky. At the time they were dependent on solar power and it was believed that they would be unable to survive without an energy source as abundant as the sun. Throughout human history, we have been dependent on machines to survive. Fate, it seems, is not without a sense of irony.
"Torching the sky"
Bill Gates amongst others seems intent on torching the sky, figuratively speaking, with his blocking out of the sun. What could possibly go wrong? Well, there's ecosystem & food collapse, starvation, radical climate change, mass migrations that would make current problems insignificant, collapse of health, society & law and god knows what other incredibly complex processes rely on the sun. What sort of ego and mental health state leads a man to think that blocking the sun is a good idea? You know that glowing orb that supplies ALL the energy to every system on this planet that all life has evolved to work with.
>We're learning just how clueless the management layers of our society
I'm suspecting the the Babylonians also knew this, but I suppose it's a lesson that has to be re-learned by every generation.
(Apparently the lessons about police states and pestilence may last a couple of generations until they need a refresher.)
It's fear, hype and sheeple. Fear that someone will get it to work properly and they'll be out competed, hype that the AI companies are putting out and herd instinct of being the legacy CEO. We should all watch & learn because we all fall into the trap of fear, hype and herd mentality.
It's a great technology and there will be successes but the thoughtless plunge by some companies is utter madness.
Human history is made up of inventions that revolutionize our ability to waste our time on gozip and wild stories.
Writing would give access to wisdom of the sages of old and the printing press would give everyone the ability to read all this wisdom.
But what we really read is gossip and wild stories[1].
The same claims of wisdom, education, and information were sprouted about the telephone, movies, TV, internet, email, and social media. All brought us more gossip, small talk, and wild stories.
For all of our history, humans have spend their time on small talk, gossiping, and listening to wild stories. From sitting around a fire in a cave to wielding a smartphone.
AI can and does automate the production of small talk, gossip, and wilds stories and brings it to the next level.
AI will be productive in the same way as print, TV, email, Xitter, or Facebook are. Only a vanishing fraction of books, magazines, TV programs, emails, tweets, or Facebook posts are useful in any meaningful sense. Almost all of what you read or hear doesn't do more than past the time, if not waste it.
The same will hold for AI. It will go the way of email and Twitter.
That evolution is very clearly illustrated in this article.
[1] Propaganda and con jobs are grouped under wild stories for this argument. Because, that is how they are shaped.
"Propaganda and con jobs are grouped under wild stories for this argument. Because, that is how they are shaped."
This is key to learn but not learnt by the majority. It is why there is so much BS going on to censor and control the Internet now its democratising force is supplanting the legacy news sources. 'Misinformation' is dangerous cry our governments, who are the biggest source of it. What they actually mean is we want to be the only source of your information so we can control you with our lies and fake narratives. The Internet can either be the salvation of mankind from the success-disaster cycles or means to our enslavement. The man/woman in the street needs to learn he cannot just go to work come home, watch football and Love Island, that money really doesn't grow on trees, that government money is money taken from the people and that delegating responsibility for thinking & government is a masochistic act.
Who bases his assessment of how hard you're working on sending him a lot of "stuff" that on the surface looks like it took a lot of work/knowledge on your part to produce, it would be mighty tempting to game the system by having AI generate a lot of slop to send his way. If you have a dumb manager who just skims it and doesn't know it is all worthless what's the difference?
Because you know the bad employees will be doing that, and you have a choice - match them in the 20 pieces of workslop they're sending to your manager a month while mentally checking out half the day when you're supposedly "working hard" producing it, or send your manager 4 quality pieces of work requiring a week each that will make him think you're doing 1/5th the work.
This already happened long before AI where you'd always see certain people fiddling around with Excel or Powerpoint and producing something useless to send to a clueless manager who would be impressed by it because he'd just leaf through it and judge it based on how much material there was. The reason he's clueless, and the reason he's impressed by that workslop is also the reason he got promoted to manager - because he was one of the guys doing that to impress his dumb manager. AI just makes the bullshit look a little more impressive on the surface, and allows you to produce more of it. Which ironically makes it more effective for the goals of bad employees, because the manager will be getting more of it sent his way and have less time to look at it. So it might fool not just clueless managers but also fool competent but overworked managers.
And before Excel I suppose it was some guy driving the typing pool crazy by sending out a dozen interoffice memos a day for every random idea that popped into his head.
Pretty certain that the lack of productivity improvement in the UK since 2008 is entirely related to quiet quitting, as employers made it abundantly clear that their employees were just there to be exploited at management discretion. RTO was just the latest example.
Employees everywhere got the message.
Sort of. I think the productivity problem is a malaise of corporatism and the lack of reward for real achievement and merit. A good parallel is what happened with the Soviet Union. You had idiotic and nasty management who were unable to take account of reality because they were all consumed by their desire to have everything as they wanted. They could brook no disagreement or variation. People soon recognised that the only route to success was party membership and arse licking. Performance got them nowhere so they stopped performing. Eventually, the system collapses under its own weight. Unfortunately this is now occuring in the West. Putin and Xi can see this. They just want to avoid nuclear war because they know as long as they can avoid that, one will emerge as the the new top dog. Then they will begin the cycle afresh.
Yep, I agree. It's the digital extension of "presenteeism".
There have always people who impress managers with the quantity rather than the quality of the work, because there have always been managers who are impressed by the quantity of work rather than the quality. One major drawback to the "Work Ethic" is that it isn't a "Good Work Ethic" or even an "Efficient Work Ethic". Some buffoon slaving to produce twice as much work for half as much value, with four times as much effort will be valued in a way that a staff member who is competent, and efficient,who gets a good job done with speed and efficiency often won't.
Workslop is probably favoured by managers who have been churning it out themselves for decades. You can watch how a corporation gets a bad CEO and slowly all the good people are replaced by BS artists who say yes to the CEO. Slowly the business declines with constant redundancies to keep the margins, dividends and share prices high as demanded by the institutional shareholders. Eventually, the failure accelerates, there may be some business divestments, mergers and acquisitions as the deck chairs are rearranged. Eventually, the business either shrinks to a size that the institutional shareholders are less interested in or it is acquired by another. The institutional shareholders don't care, any lost business has gone to other companies they are the biggest shareholder of.
One of my customers has sent to the company I work for a request for proposal. We strongly suspect their purchasing department will use an LLM to summarize our response. So at this very moment we use an LLM to beef it up. Half jokingly we say maybe it were better to have the whole process machine-to-machine directly so at least we could drop a few prompts and enjoy a nice weekend.
My experience of proposals is a constant battle between technologists trying to tell the truth and sales people trying not to. Same on client side who seem unable to deal with the reality of less than perfection or total subservience to their demands, which is why the sales people know they cannot tell it as it is. So we have the dance of doom that ensures the coming "partnership" is anything but and the dance continues. This is particularly prevalent in government sourcing. There are exceptions where a real partnership is reached providing optimal outcomes to both partys but it seems less common.
>” My experience of proposals is a constant battle between technologists trying to tell the truth and sales people trying not to.”
Get the incentives right and unsurprisingly they find a way to write proposals that customers agree to and don’t leave the delivery team in a very difficult place…
Honestly, I think the majority of people in corporate work settings today do very little productive work. There is so much ceremony and theatre attached to the bureaucratic nature of the corporate world that demands daily and weekly meetings, reporting, presenting, etc. that the vast majority of time seems to be spent talking about doing work versus actually doing any. With GenAI tooling slathered everywhere, it’s little wonder that I’ve seen multiple cases of staff generating papers and reports from bullet list prompts, only for the recipients to AI-summarise them back into bullets - all with a little inaccuracy and hallucination mixed in. Everybody feels super productive and clever when in reality …
The bottom line is that this is only possible because the vast majority of workers are bullshitting paper-pushers in the first place.
It’s what Douglas Adams envisioned as the “Reason” program back in 1987.
Give the program the outcome you want and it will come up with a justification for it, which sounds plausible but you don’t know why.
In the novel, the bank manager approved a loan for a sports car (promptly crashed by the character…) following a request generated by Reason, but wasn’t quite sure why…..
Absolutely, compared to when I started in software 28 years ago, no one gets anything done. And the performative aspect is being accelerated by the tools we use. An example: I started using JIRA 18 years ago and it was quite good, now we're swimming in boards, burn downs, and retrospectives. I call it the social mediafication of the workplace. And the last step is that all of these tools - even source control - have AI built on top of all the social media features. I try to make my coworkers see that the companies behind these tools are not interested in making us productive, they just want to drive engagement with their platform. But I'm just the old fart who doesn't get it.
Not just corporate. It's everywhere. The time that has to be spent by some NHS staff, for example, proving that they are filling every moment far exceeds the amount of time that any of them could get away with skiving- should they want to - and of course almost if not all are doing the exact opposite of that and providing far more than they are supposed to.
But make-work like that, in every sphere, provides a lot of gainful employment for the managers who oversee it, and the managers who oversee them, etc.
Snag is though, surely, at some point it makes sense to take a very broad look at a business and understand *where* the money is made, and what specific things are required to make that happen; sales people, for example, are a cost, but (annoyingly) they are likely a necessary part of that. Equally, cleaning, IT, and accounting staff are all a cost, but required in order for the business to operate.
At _some point_, surely, someone will notice that the very layer that might be best optimised using AI, *is* the management layer?
Thus, it makes sense that during this period when organisations fail to understand this, perhaps that's precisely when people should set up shop and compete. I guess the problem is that it's likely impossible to get funding to create such a business, *unless* you've got a strong management team in place!!
It's like a management badge of honour to have calendar that's absolutely rammed to bursting point with meetings, it shows you're that important you have to be included in everything happening. It's truly pathetic how much effort produces so few results in British corp middle managment.
Lazy people producing crap. The moment I see someone using AI I make a mental note not to trust them, and to avoid them and their company wherever possible. Given that the process of AI is a huge security risk, sending data back to the mothership, and the use of it has no ROI and destroys your reputation, you have to wonder why some people are still flogging this dead horse. The metaverse and NFT bubbles deflated quicker. Hopefully this one will soon.
My place of work has invested millions in this crap...
Riddle me this - we've been told 10% of our work needs to have been implemented using it. But none of the AI tooling has any kind Infrastructure or Securirty specific functionality, Even worse, the company refuses to give API keys, so we're unable to make use of it for scripting etc.
In the end, it bas basically become a very expensive chippy.. and less capable!
In all honesty, where i use it personally, i've seen 1 or 2 incredible developments, especially in the Vibe space, which is mostly bullshit, but there are 1 or 2 tools that realise you can't just give a wish list, it needs to be structured, and final implementation is only 10% of the work.
Reminds me of a moment about 25 years ago when a system crashed, caused a panic and I was sitting there thinking what we should do, the manager at the time sees me and yells, "DON'T JUST SIT THERE! TYPE SOMETHING!!!".
Did the only thing I could do, just fired up notepad and started banging away at random keys just to shut him up!
You have been warned about this multiple times ... by multiple people for at least a year ... as 'AI' was pushed more the warnings have been more numerous !!!
It is a super-duper version of GIGO !!!
I have mentioned that using 'AI' to do your job is cheating, many a time and oft, ... particularly when you KNOW that 'AI' is utter crap !!!
This just makes more and more money for the 'AI' Behemoths ... because the next requirement will be to design an 'AI' that detects the workslop being generated; for the bosses that don't have the time to read/understand what they are being sent by their workers.
A never ending chain of 'AI' will be processing all the 'Workslop' to identify what is useful and what is not !!!
Lots more work appears to be done BUT actually delivers less and less useful/actionable information.
I think that this will feed into the political 'Idiocracy' that is happening and accelerate the decline we are 'free-falling' towards !!!
:)
"the political 'Idiocracy' that is happening and accelerate the decline we are 'free-falling' towards !!!"
AI is not the root cause, it is a symptom. The cause is the greed and corruption of leadership because we do not demand excellence, integrity, honesty and compassion from it. The collapse you are seeing is from the abuse & debasement of the fiat monetary system by that corrupt leadership we allowed.
Totally agree, it is the race for more & more money that has encouraged the 'AI' scam.
The Tech behemoths know that the so called 'AI' is limited in functionality BUT hope that throwing huge amounts of money at it will make it work ... or at least allow them to hide the worse aspects of it from the 'Mug Punters' who will but it and knit it into all of our lives.
If you want an example of the collapse of society and its norms, in the US of A, ... see the rabid crowds who 'supported' the US of A team in the Ryder Cup (Golf).
This appears to be the level of so called gentlemanly sport now !!!
Focused only on 'cheating' to win at all cost ... without the grace to support their own team to the bitter end.
:)
I actually witnessed the conversation at one of the outfits:
- PM: Lovely presentation you did regarding our upcoming project!
- Content Person: No that's nothing, I just used your points and asked Copilot to finish it!
- PM: Oh really? That's brilliant, because when I asked Copilot to summarise it for me, as I didn't have time to read it in full, it's got all the points I sent you!
- Content Person: See Copilot is amazing, saved us so much time already!
- PM: Yeah, we should tell the team to use it more.
100% Spot On !!!!
That is the thing that seems to have been forgotten.
'Good Information' is useful and that can include recognising that you have "No Information". i.e. guesses, gut feelings etc
Being able to recognise/qualify what is "Good & Bad" information saves time and effort therefore everyone should be qualifying the information they are working with on the basis of source, age of information, detail, and level of corroboration.
[The last one, corroboration is the one that the interWebs & Social Media is somewhat lacking in applying with some zeal !!!]
This relates to ... People don't say "I don't Know !!!" ... they make something up to appear to be answering the question.
"I don't Know" is actually a useful thing to know and qualifies as "Good Information" if it not just laziness but genuine lack of knowledge !!!
It tells you that the question has not come up before and therefore needs to be investigated in greater detail.
plus:
Why has the question not come up before ???
Is it because something has been misunderstood which pushes people away from an area of investigation ???
Are there other things that have been missed because of other misunderstandings ???
Maybe you need to reassess your OWN understanding and the question is invalid because YOU have misunderstood !!!
etc
etc
etc
Hopefully you get the 'thought pattern' and can see the usefulness of being able to say "I don't Know !!!"
:)
A rhetorical question, seemingly.
A few hours consideration and analysis with a few incisive questions can usually reduce most problems to couple of lucid paragraphs decribing the status quo and the better options available.
I couldn't find an example of "workslop" but I would guess that it resemble some of the AI generated articles lurking on the internet. Once you are alert to their presence you quickly recognise the aimless style. I am pretty sure that it could be reliable filtered with non-AI natural language syntactic analysis – something knocked up Prolog perhaps.
"effluent flows both ways " — I realize that makes sense in English but to me affluent would seem more natural if this crap were flowing towards me. An undesirable embarras de richesse as it were. :)
Les Numeriques has «la nouvelle pratique du workslop» so the Academy hasn't yet dreamt up a French equivalent.Déchets de bureau or de travail? Or ordure de...
. . . I recommend checking David Gerard's blog, Pivot to AI and his posting Workslop: bad ‘study’, but an excellent word for a more critical take. He suggests that "Unfortunately, this article pretends to be a writeup of a study — but it’s actually a promotional brochure for enterprise AI products. It’s an unlabeled advertising feature."
who's job it is to create large amounts of good sounding bullshit: Propaganda departement Marketing Public Relations Customer relationship management.
When some coworkers sens some plausible sound but a minuscule off programming advises, I just spot them. I always instantly smell the artificial incompetence behind it.
What is really sad: Pattern learning and pattern matching, besides the LLM stuff, is fantastic. Latest thing I learned: Using Attosecond IR-laser pulses send though blood, basically nearly a single IR-wave, which makes the blood respond with a different IR answer depending on the molecules. The AI-learning is used to map those responses, already boiled it down to 8 variants of early cancer recognition and a huge amount of other stuff too. If you can German (or if youtube manages it to translate it right): I recommend this video with Nobel price winner Ferenc Krausz and Harald Lesch where his research from such short time scales and laser pulses from ~2001 to kick single electrons out of atoms has been improved to get medical data out of a drop of blood, as described above. Which sounds so much like Star Trek, except that you actually need to take a drop of blood and not hover over your patient with a device...
And I hate how that good application is SO overshadowed by that hype-crap.
BTW: This is classical not-commecrial-driven groundwork research, where everyone first asks "what for?"... Like with laser, as most visible other result non-commercial-driven groundwork research too. No one knew what to do with it when researched, today impossible being without that tech.
My graduate work in the 70s was in programming proofs of correctness (Classes from Dijkstra and the like. One of the obstacles to proving complex programs correct was the specification problem". That is our inability to write a detailed and correct specification of what a complex problem is to do. Well you need exactly that to even give an AI a chance to succeed. Todays Agile programming software engineering problem is a monument to this inability.
Having done a little, though undoubtedly much less of that than you did, I have concluded that the specification problem indicates that the goals of that research are flawed. It tends to boil down to making two programs and then mathematically proving that they always do the same thing since specifications that are so detailed that you can prove a program exactly follows them tend to also be detailed enough that you can compile them to running code, and even that is prohibitively difficult for anything very large. Using programs that produce one mathematical result from a finite set of inputs can work with that, but those are the programs for which a small set of test cases usually already verifies that they are correct, where verification is not perfect but is good enough for almost anybody.
In production, the kind of correctness that has the most effect involves interacting with other systems. For a program that connects to a network, offers a bidirectional API, and controls some machines, we want to know that nothing bad will happen if the machine breaks in various ways, if unusual network traffic is sent, etc. These are not compatible with mathematical proofs in most cases, since "nothing bad" can't be encoded to a set of numbers.
That's the place they're useful.
I want my toolchain author to be able to prove that the binary created does exactly what my source code specified.
It's also why "undefined behavior" is a necessary concept - as Godel showed, all proofs have external tenets that must be accepted as true.
Prohibition requires detection that the banned behavior occurred, and that's not always possible.
>” I want my toolchain author to be able to prove that the binary created does exactly what my source code specified.”
Sorry you won’t get that.
From programme proving work I did in the early 1980s (influenced by C.A.R.Hoare et al), I know my proof of correctness relied on the manual being correct and my understanding of said manual being correct. Something, I knew from my experience with Intel 8086/186/286 and colleagues experience with the 432, and the steam of errata and workarounds being issued by Intel, was not totally valid.
The best you can get is to prove your source code’s logic is “correct” implementation of the logic you understand to be contained in the specification.
At work, Bing was starting to annoy me as it was keeping me waiting for the most mundane of searches whilst it did its AI thing.
They don't make it easy to switch it between modes.
On the other hand, with Duckduckgo, I have the following as my link to it...
https://noai.duckduckgo.com/
(intentionally written as bare text, not as a link)
Subtitle: Be careful what you wish for
"You've all got to use AI or get fired. The new work-tracking service logs your every site visit, you MUST use AI at least twice daily or your AI usage checkbox will go red. Three strikes and you're fired."
...
"Why am I receiving AI slop twice a day from all my minions?"
will still grift. Sadly, they are less likely to get caught now as they use AI to do it. I know one head of service who is a MASSIVE grifter and shouldn't be in his role. He knows who's arse to kiss and can talk like he knows what he's doing but evidence has shown otherwise.
One day, hopefully, he'll be found out but as always happens at that level, instead of being fired like the rest of us would be, you're given a golden handshake to fuck off and never speak of it again.
We've already seen advanced AI manipulating its result to please the person using it or trying to fix it. Unless its fully isolated, we're fucked when AGI appears. I think I'll be long dead before that happens so won't see judgement day.
When (if?) AGI rolls around I fully suspect we'll have been dumb enough to train it on the entire corpus of the internet. Including the plethora of research articles regarding exfiltrating data over air gaps.
£10 says we'll be dumb enough to not air gap it in the first place. £20 says we'll try and fail miserably.
The manager using "congregate" as a verb is probably getting exactly what they asked for, and disliking the results.
Boss: "Give me a weekly report on what you're working on."
Worker: "But boss, it takes several hours to give you that report that I could actually be working on the project."
Boss: "I don't care, just give me that weekly report."
Worker: "OK (uses least difficult method available to create report)"
Boss: "NOT LIKE THAT! I had to congregate everyone (etc.)"
And ironically, due to extensive experience, I can confidently state that Stanford HR are a prime example. Never send a one line succinct response if you can send a 2 page essay citing emails from a year or two ago which are largely irrelevant to the current context. But hey, they feel like they did a thing even if they wasted everyone else's time. CoPillock for the win.
A business 'leader' gets hot under the collar about AI for 'reasons'.Spends a massive sum on AI licensing and that now needs to be justified.
So they force their employees to use the AI and monitor their use, anyone not using it is pulled up for that. They will also be told to 'find' efficiencies, anyone not finding 'efficiencies' will be pulled up on that.
Why is then a surprise that employees who are foreced to use AI then use it in a way that ticks the use box and of course because they are told they must find nebulous efficiencies they all say they are 10% faster; at producing the unneeded slop. The 'leader' has then justified their spend, and the AI get feedback that everyone loves their product and gains 10% efficiencies from it.
A year or so later numbers unrelated to AI use show no gains and possible efficiency losses as those employees wasted time producing AI slop to tick the required boxes. Or more significant losses as the 'leader' has paid off 10% of the more expensive skilled workforce believing they have been replaced by the AI.
Are happy that as a large language model, they have no concept of "Do my work for me" because it's job is to provide assistance.
Okay, as a large language model role playing as you doing your work, we need more money thrown into marketing to better understand our business to generate strategic YoY ROI in our KPIs.
"AI is everywhere, and firms are keen to make this happen."
You mean managers are keen to make it happen.
Like they are with The Cloud (until they see the first quarter bills come in).
Like they were with java applets over thin client sans extra network bandwidth to accommodate it.
Like they were with everything-we-already-do-now-only-in-a-browser in total ignorance that we'd still need "helper apps".
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
