One in six US workers pretends to use AI to please the bosses If you're one of those people who pretend to use AI at work, then worry not: there are likely another 15 of you per hundred employees in your company. That's the finding of a survey from nearshoring tech recruitment company Howdy.com. According to the survey, three-quarters of people's employers expect them to use AI in some …
The top management are absolutely drunk on the stuff, seeing $$$ return in using it.
As it will be detrimental to my ongoing employment at the company, pretending to use AI is exactly what I do.
One of our AI evangelists said this recently...
"You may not lose you job to AI,
But you may lose your job to an Engineer who uses AI"
"Twice now I’ve had cow-orkers contribute to a “what does this part of the standard mean” conversation"
I'm more impressed that you've got cattle where you work (doing some "orking" apparently) and that they've contributed to a conversation not once, but twice.
Apologies, but I couldn't resist.
This post has been deleted by its author
There is no I in AI. It's painfully obvious when someone uses a chatbot to compose an email or letter. It's painfully obvious when they use it to summarize another document. It's painfully obvious when someone uses it to make an image or video. The only people it fools are other dunces.
A"I" is a tulip of potentially historic proportions.
I'm not even sure if it's for Copilot on Windows or Copilot on Github or why the company wants to spend money on this nonsense. Nevertheless applying is the thing we're told we have to do, so I'll apply for it then carry on as before.
A presentation where we were shown the AI plug-in in Visual Code Studio "hallucinate" code in a programming language we don't currently use was deemed a resounding success by management. That's the kind of nonsense that is impossible to fight against with logic or reason, so I'm just going to tick the box and carry on.
I've no idea who invented that usage of the term, but "hallucinate" was always a poor analogy both in terms of what it *looks* like is happening *and* what is "actually" happening under the surface.
On the other hand, while "bullshitting" is probably closer to describing the surface *appearance* of such behaviour, it's still too close to anthropomorphisation for my taste. It implies some level of awareness- or at least the *ability* to be aware- that they're bullshitting. If a human did this, that would be a fair assumption, but that's not really how LLMs work.
OTOH, I'm not sure what we *should* call it.
Unfortunately, "lie" has the same problem as "bullshit" if the machine doesn't "know" it's lying!
"Wrong" can definitely be applied, because it only applies to the end result- it doesn't ascribe any underlying motive or behind-the-scenes explanation.
I work in the public sector, it's happening here too. Whilst it's not being demanded as such, because the bigwigs (from the very top, ie politicians) think AI is so wonderful, there's a lot of soft pressure to apply it. I've taken the King's shilling, so I've got myself a CoPilot licence (and across all of O365 it's a weird collection of different interfaces, different responses to the same questions, refusal to do simple, obvious stuff in one app, and do it elsewhere). Has it made my output better or quicker? You know the answer already. A few times its saved me half an hour by doing low value stuff like summarising things where it actually doesn't matter if the summary is not correct, or a first draft of meeting notes, but so what? Anything actually valuable? Nope.
Wonder how much MS invested to produce and deploy a tool of such staggering un-usefulness?
One has to wonder how the Management in your company was convinced that the way to 'use' 'AI' was to mandate that everyone registers for a CoPilot License.
Useful tools are used because they are 'USEFUL' not because someone is holding a gun to your head.
MS must be overjoyed with the number of CoPilot licenses your company has obtained !!!
Is there any evidence of 'AI' being used for anything that the company is gaining from ???
I am stunned by this BUT well done to the MS Sales Drone who sold this idea to your Management !!!
:)
I told him, not really,
Anything that I need help with, the bot would probably not have a useful answer for
And anything that it would be useful for I can do faster by myself.
It is useful to those who have little experience on what the AI has been trained on.
It's just an easier interface into KB articles and other documentation.
All it does is the equivalent of generating as many propositional statements from a piece of documentation,
Generating a list of propositional statements describing someones AI chat query requirements.
And finding the best overlaps.
We are a support organisation providing support for a number of applications so restricting responses to those that are most appropriate to the team we are on is also done as a heuristic.
No need for documentation on application abc when you are in a team supporting application def so pull in data like that to focus the AI response.
Bosses just want AI to replace the need for SMEs to get the job done. They can afford to pay a lot less if they can replace competence and experience with button monkeys who just need to understand the AI response enough to prevent it from telling customers things that make no sense at all. They just want a sanity check on AI responses and to hide the fact from customers that they are being primarily served by AIs that do not understand their requirements at all.
>We are a support organisation providing support for a number of applicatins<
Your boss should be glad that AI isn't better or as good as you are on that job, If it were your clients would replace your organization by an AI. The customers might well do that before your boss makes back the money invested to replace you.
Where I work we have a company-specific Chat-GPT instance, and the business is monitoring for anyone who has not logged into, and used it. I haven't because it's a pile of shite for most applications (never mind that the power requirements are eventually going to burn the planet). Being so close to retirement, I'm looking forward to any chats about my lack of compliance.
I see "AI" as a tool. When I need that particular tool I will use it. In the meantime, the tools I have are doing the job just fine. I have tested the use of AI in support of what I do, and whilst one AI produced data faster and more efficiently in one specific area, another AI hallucinated stuff that does not exist and generated outright bullshit. A decent search engine can do the former.
This whole "everyone must use the company AI" has the same stink as all these return-to-office mandates. Somebody invested too much money/ego into it and is determined that everyone see what a great idea it was of theirs. What is going to be really funny is watching the panic amongst the "AI" promoters and advocates when the bubble bursts.
Your description is the exact thing that makes current 'AI' so useless !!!
No matter how Good or Bad the 'AI' can be, the real issue is that you have no idea if the interaction TODAY is the one that will be Good or Bad !!!
The fact that you don't know if you will be getting a useful answer or not means that you cannot rely on using 'AI' at any time.
If you are in a hurry and need a quick AND accurate answer NOW ... it will be a roll of the dice what you will actually get !!!
IF 'AI' is effectively a random 'Answer Machine' then you could just as well use a 'Magic 8 Ball' to 'guess' the answer !!!
:)
If management is completely f@#$ing off the rails stupid and incompetent and demanding you use 'AI' where it is actively detrimental and unsafe, then malicious non-compliance is 100% justified.
Management are not your friends and are actively malicious and hate you, so treat them as the enemy they are.
> You must have worked for some shitty companies.
Any corporation traded on a public stock exchange would literally grind live puppies, kittens, and duckies (even human infants if you care about those annoying things) into blood meal if it were overall profitable enough. And they would claim it was a moral mandate for stockholder value. So yes, at one point I worked for a corporation. Never again.
TO BE FAIR, since then I have worked for small companies and management have not been actively malicious (mostly decent or at least harmless) and I would not actively try to sabotage them in return. But mandating you use AI is proof of maliciousness, so if they go there then feel free to retaliate.
The company should support the service, but the onus to show up, practice, learn, be curious, and work at it is on the individual,
No!
The onus of the employee is to turn up and do the work they are paid for. If you want them to learn a program, AI, anything, you damn well pay them for it. Make sure they have the time on company time to do the training and practice with it. If that means they can't do their regular work that's on you.
It's not up to employee to do anything but the job they're paid for...
"Jacqueline Samira, CEO at Howdy.com, says [...] employees also need to embrace the new tech that's coming down the pipe"
Why, specifically? Some things that come down some pipes are definitely better not "embraced".
BTW, I've just marketed a snodfangle grunging grabber, so everyone must adopt it (whether or not it's actually useful).
Company is very keen on embracing AI.
It is occasionally useful (in my experience, referring to github copilot as dev related use) only on some relatively straightforward tasks - and even then, to get results that need minimal tweaking you need to be very detailed and explicit on the prompt you give (which takes more time, so soon risking it taking nearly as long to develop a prompt as it does to create the code yourself).
Anything complex (beyond real boilerplate stuff) has so many errors to correct that it is a total pain... The obvious errors are not too bad, it's the subtle, more difficult to spot ones that I despise.
It is OK on boilerplate unit test code, once you have written a couple in a consistent style it soon gets the hang of creating those with relative concise prompts & so can be significantly faster than DIY coding.
Currently only of benefit in limited use cases (& played around with it enough to know when to use it & when not)
There are some more advanced AI coding tools gradually getting pushed out, so will have to see how those go.
Except that China is throwing around $90 billion at AI on its own account this year alone. The West is ramping up for a re-run of the dotcom bubble, and China will be joining the party with billions of unproductive investments (to perhaps add to a range of unproductive housing and infrastructure investments that still have to be fully unwound).
Lol. The release of Deepseek traumatized the entire Silly Con Valley into alternating suicidal despair or rabid nationalism for all of five minutes until their pets in the government banned it.
The reds know what they're doing.
The question is is anybody in power in the West not too busy sucking at the American style, neoliberal workfare-for-MBAs ("We can cut so much more environmental and labor law! Line go up! AI gonna save us! Coal gonna save us! Vote for me!") think-tank tit to notice the tides changing?
I've seen this. This isn't the tech savvy boss who wants to replace the workforce with robo drones. This is the lazy boomer boss who never really learned computers, and doesn't realize when they write emails with chat GPT, their employees clean up after them, so they think that chat GPT works way better than it does.
