All IT Work To Involve AI By 2030, Says Gartner (theregister.com) 61
An anonymous reader quotes a report from The Register: All work in IT departments will be done with the help of AI by 2030, according to analyst firm Gartner, which thinks massive job losses won't result. Speaking during the keynote address of the firm's Symposium event in Australia today, VP analyst Alicia Mullery said 81 percent of work is currently done by humans acting alone without AI assistance. Five years from now Gartner believes 75 percent of IT work will be human activity augmented by AI, with the remainder performed by bots alone.
Distinguished VP analyst Daryl Plummer said this shift will mean IT departments gain labor capacity and will need to show they deserve to keep it. "You never want to look like you have too many people," he advised, before suggesting technology leaders consult with peers elsewhere in a business to identify value-adding opportunities IT departments can execute. Plummer said Gartner doesn't foresee an "AI jobs bloodbath" in IT or other industries for at least five years, adding that just one percent of job losses today are attributable to AI. He and Mullery did predict a reduction in entry-level jobs, as AI lets senior staff tackle work they would once have assigned to juniors.
The two analysts also forecast that businesses will struggle to implement AI effectively, because the costs of running AI workloads balloon. ERP, Plummer said, has straightforward up-front costs: You pay to license and implement it, then to train people so they can use it. AI needs that same initial investment but few organizations can keep up with AI vendors' pace of innovation. Adopting AI therefore creates a requirement for near-constant exploration of use cases and subsequent retraining. Plummer said orgs that adopt AI should expect to uncover 10 unanticipated ancillary costs, among them the need to acquire new datasets, and the costs of managing multiple models. The need to use one AI model to check the output of others -- a necessary step to verify accuracy -- is another cost to consider. AI's hidden costs mean Gartner believes 65 percent of CIOs aren't breaking even on AI investments.
Distinguished VP analyst Daryl Plummer said this shift will mean IT departments gain labor capacity and will need to show they deserve to keep it. "You never want to look like you have too many people," he advised, before suggesting technology leaders consult with peers elsewhere in a business to identify value-adding opportunities IT departments can execute. Plummer said Gartner doesn't foresee an "AI jobs bloodbath" in IT or other industries for at least five years, adding that just one percent of job losses today are attributable to AI. He and Mullery did predict a reduction in entry-level jobs, as AI lets senior staff tackle work they would once have assigned to juniors.
The two analysts also forecast that businesses will struggle to implement AI effectively, because the costs of running AI workloads balloon. ERP, Plummer said, has straightforward up-front costs: You pay to license and implement it, then to train people so they can use it. AI needs that same initial investment but few organizations can keep up with AI vendors' pace of innovation. Adopting AI therefore creates a requirement for near-constant exploration of use cases and subsequent retraining. Plummer said orgs that adopt AI should expect to uncover 10 unanticipated ancillary costs, among them the need to acquire new datasets, and the costs of managing multiple models. The need to use one AI model to check the output of others -- a necessary step to verify accuracy -- is another cost to consider. AI's hidden costs mean Gartner believes 65 percent of CIOs aren't breaking even on AI investments.
I mean, it already is to some degree (Score:4, Insightful)
Oh, they meant "genAI". That's probably going to collapse by the end of the year.
Re: (Score:2)
Yes, that's the garbage (Score:2)
That won't make it past the end of the year.
Re: (Score:2)
Yep. Or maybe next year if they can keep the stupid excited a bit longer. But there is no way this crap will be profitable anytime this decade.
Re: (Score:2)
I'm one of the stupid then. I use AI daily, and find that it increases my own productivity significantly. Mainly, it helps me avoid having to look up stuff in reference pages.
OK, not genAI, just AI, specifically GitHub Copilot.
Re: I mean, it already is to some degree (Score:1)
Re: (Score:2)
Well, I probably shouldn't dignify your post with a response, but I'm going to anyway.
Your statement about management has nothing, NOTHING, to do with AI. AI doesn't make bad coders good. AI makes good coders, more productive. It does the nit-picky drudge work they don't want to do, that slow down the process of bringing code to life.
As for profitability, *every* new technology is expensive in its early days. Then with each iteration, smart people figure out ways to accomplish the same goals, more cheaply.
Re: (Score:2)
And what does that have to do with the tools being profitable?
Re: (Score:2)
Well, sure, I addressed your first sentence.
As for profitability, every new technology is expensive in its early phases. And then smart people come along and find ways to make it better and cheaper. This is especially true in the area of electronics and software. The original base model IBM PC cost $1,500, or $5,200 adjusted for inflation. That machine was peanuts in terms of performance and capability, compared to today's base model laptops, which start at about $300.
AI will follow the same price and perfo
Re: I mean, it already is to some degree (Score:2)
Re: (Score:2)
It sounds like you're using the web version of ChatGPT or Copilot or similar. You might want to consider GitHub Copilot (Visual Studio, Visual Studio Code), or some other AI integrated with an IDE. It is aware of your context, and will customize the "sample" for you, plugging in the right variable names automatically. Or you can right-click some lines of code and ask it to refactor them using a specified strategy. Very handy!
Re: (Score:2)
Thanks for the tip
However, I'm really quite leery of handing my code and thought processes to Microsoft, even though I do have a github account... I used it less and less after microsoft bought it, seems like a long time ago now. Fortunately, I don't have bosses, so I can make that decision.
I actually found a web proxy site to access the bottom tier LLM's which a couple Russian graphic artists based in the UK, threw together to get between you and Big Brain
If I wanted the wrong solution, my manager (Score:3)
Re: (Score:3)
Re: (Score:3)
And it may prove to be the most prophetic of all ST:TOS episodes. Unfortunately, the C-suite is currently jacking off to it. "No, don't let that pointy-eared freak control the guard's mind! And pass my cocaine!"
Re:If I wanted the wrong solution, my manager (Score:4, Interesting)
If the wrong solution was good enough
If you ask the likes of the Klarna or SalesForce CEOs, the wrong solution is already good enough. The greatest decline in software quality has only just begun. Decades ago we used to think that badly optimized code was the worst that could happen if poorly educated people did the programming, then it was unstable code that became a "usual" outcome, but now we have reached the point where "software that does not quite perform its primary function" has become an expected result from the LLM wielding vibe-coders that are hired. And I am horrified to observe how "Product Managers" already accept this as the "new normal", and ready for shipping.
Re: If I wanted the wrong solution, my manager (Score:1)
+5 well put
Re: (Score:2)
If the wrong solution was good enough
If you ask the likes of the Klarna or SalesForce CEOs, the wrong solution is already good enough. The greatest decline in software quality has only just begun. Decades ago we used to think that badly optimized code was the worst that could happen if poorly educated people did the programming, then it was unstable code that became a "usual" outcome, but now we have reached the point where "software that does not quite perform its primary function" has become an expected result from the LLM wielding vibe-coders that are hired. And I am horrified to observe how "Product Managers" already accept this as the "new normal", and ready for shipping.
Agreed. However, blaming AI alone for the acceptance of "barely functions" software as shippable product is a bit off-base. Accepting the output of AI whether it meets the spec or not and whether it's bug-filled or not is just a slight moving of the needle on a problem that has grown exponentially since the internet / web became an expected commodity for the end-consumer. Once companies realized they could ship completely broken programs and fix it with a service pack, or other form of online update, we've
Might be possible... (Score:5, Insightful)
But broadly speaking it's felt like throughout my entire career Gartner has said various things with all the accuracy of coin flipping. I'm shocked that business people have kept citing them time and time again like some grand Oracle as they keep flubbing the details with little or no particular insight than anyone else.
Re:Might be possible... (Score:5, Funny)
Gartner has said various things with all the accuracy of coin flipping.
I disagree. Gartner has been very reliable over the decades. I have routinely done the opposite of Gartner's predictions, and come out ahead 90% of the time.
Re: (Score:2)
Mod parent funny!
Paid for in divorces, broken homes, ... (Score:2)
Gartner, McKinsey and Booz Allen Hamilton, all business consulting strategy firms advised US corporations that offshoring was an effective long term business strategy.
It resulted in tens of thousands of layoffs, with every thousand laid off having a few end in divorce, broken homes, addiction and suicide.
From 2011, National Institute of Health - https://pmc.ncbi.nlm.nih.gov/a... [nih.gov]
" using job histories of workers in Pennsylvania, Sullivan and von Wachter (2009) find that mortality risk doubles in the year of d
Re: Might be possible... (Score:2)
You didn't get it wrong, they did.
Re: Why not (Score:2)
thinks massive job losses won't result?!?!? (Score:3)
The effort is worth something - to the company, if it was just a break even, it wouldn't happen. Even the increased productivity gets sucked up by ownership / shareholders. That's what's happened since the 1970s with technology and there's zero reason to expect it to change. https://www.epi.org/productivi... [epi.org]
Eventually there will be only 2 jobs left.. the guy who owns everything and the other guy who cleans the 1st guy's toilet. Because, even robots have standards.
Anything that can be automated will be automated (Score:2)
Re:Anything that can be automated will be automate (Score:5, Informative)
If you find yourself doing repetitive things every day that's AI can almost do, it's time to look for another career.
Most of that stuff has already been automated, without the assistance of AI.
Re: (Score:2)
Yep. It is called "libraries", "frameworks" and sometimes even "code generators".
Re: (Score:2)
Re: (Score:2)
Well, as these are probably people doing "vibe coding", it matters little. Whatever they produce will be crap anyways.
Re: (Score:2)
Excellent (Score:2)
This just in: Gartner predictions unreliable (Score:4)
These people like to hallucinate when they think it will get them money.
Re: What's next? (Score:2)
Gartner doesn't know anything (Score:2)
all of it is true! (Score:2)
It needs some work (Score:2)
I was just asking chatgpt for help with an error message and part of the fix it recommended was to rename a file to itself. (mv X X)
You might successfully use it for summaries (Score:2)
You can use it to summarize, but you still need old-fashioned non-AI detection logic for critical monitors. Will everything be AI-augmented? Probably, alas. Should it be? For lots of it, no, for some of it yes. Should it replace anything? I doubt the hell out of that. Additional reports in the form of LLM-generated summaries are one thing. Depending on it to do any job that is important? That's a paddlin'.
AIpocalypse Now (Score:2)
It won't take until 2030.
Everytime name resolution gets screwed up, I start a ChatGPT session "help me un**ck systemd resolved".
At this point, ChatGPT is like "Point to where resolved hurt you." Smart ass.
But I would spend 3x as long fixing my configuration without it.
Re: AIpocalypse Now (Score:2)
By choice? (Score:2)
You mean today...? (Score:2)
2030... what? I work in IT, I roam around IT companies, and I haven't seen one where AI isn't *heavily* involved. NOW.
Boring... (Score:2)
Remember when all movies would become 3D? (Score:2)
GenAI is already in decline (Score:2)
With the latest METR and MIT studies showing no productivity benefits (and even declines), the technology getting more expensive every day (despite what companies say about inference costs being lower) because of diminishing returns to improve model quality, GenAI adoption is already in decline.
This could have some credibility one or two years ago. Nowadays it is another junk prediction in the line of the infamous AI 2027 report.
Wont happen. (Score:2)
It wont happen. 2030 is the milestone whereby demand for clean water outstrips supply.
Ergo - if were aiming to scale AI up to that level by 2030, that situation can only worsen , not improve.
Clean Water is a very important component of cooling datacenters.
Denial continues.
Re: (Score:2)
It wont happen. 2030 is the milestone whereby demand for clean water outstrips supply.
Ergo - if were aiming to scale AI up to that level by 2030, that situation can only worsen , not improve. Clean Water is a very important component of cooling datacenters.
Denial continues.
The solution to that problem is obvious. Take away drinking water from poor communities to keep the datacenters growing. Duh.
Yes, this time we are the horse. (Score:2)
We all know everyone will be replaced the day it will be cheaper or better to do so. There is no point at raging at reality or trying to stop the Singularity.
for now make sure it is not worth it to replace you, this means all the processes at your company only work as long as you are there.
My alt take (Score:2)
My alternative prediction: no IT work will be possible except in airgapped environments because AI malware will make having a device connected to the internet essentially impossible. Site visits will be a thing again.
Good Riddance On Those Entry Level Chuds (Score:2)
I mean, it's not like we're ever gonna age out, right?