AI Could Affect 40% of Jobs and Widen Inequality Between Nations, UN Warns (cnbc.com) 40
An anonymous reader shares a report: AI is projected to reach $4.8 trillion in market value by 2033, but the technology's benefits remain highly concentrated, according to the U.N. Trade and Development agency. In a report released on Thursday, UNCTAD said the AI market cap would roughly equate to the size of Germany's economy, with the technology offering productivity gains and driving digital transformation. However, the agency also raised concerns about automation and job displacement, warning that AI could affect 40% of jobs worldwide. On top of that, AI is not inherently inclusive, meaning the economic gains from the tech remain "highly concentrated," the report added.
"The benefits of AI-driven automation often favour capital over labour, which could widen inequality and reduce the competitive advantage of low-cost labour in developing economies," it said. The potential for AI to cause unemployment and inequality is a long-standing concern, with the IMF making similar warnings over a year ago. In January, The World Economic Forum released findings that as many as 41% of employers were planning on downsizing their staff in areas where AI could replicate them. However, the UNCTAD report also highlights inequalities between nations, with U.N. data showing that 40% of global corporate research and development spending in AI is concentrated among just 100 firms, mainly those in the U.S. and China.
"The benefits of AI-driven automation often favour capital over labour, which could widen inequality and reduce the competitive advantage of low-cost labour in developing economies," it said. The potential for AI to cause unemployment and inequality is a long-standing concern, with the IMF making similar warnings over a year ago. In January, The World Economic Forum released findings that as many as 41% of employers were planning on downsizing their staff in areas where AI could replicate them. However, the UNCTAD report also highlights inequalities between nations, with U.N. data showing that 40% of global corporate research and development spending in AI is concentrated among just 100 firms, mainly those in the U.S. and China.
Everyone get on the bandwagon (Score:3)
AI is projected to reach $4.8 trillion in market value by 2033
Based on what? Altman's greed?
Re: (Score:2)
Altman controls a pretty tiny fraction of the market.
Re: (Score:2)
Which "market" is that? The market for fake degree certificates, the market for fake scientific papers or the market for fake news?
Re: (Score:2)
The "AI" market starts from TPU/GPU development, goes to datacenters made to house them for the purposes of training and inference, and the products built off of them.
The work done by them right now include medical research, physics research, customer service, and development.
Re: (Score:3)
The "AI" market starts from TPU/GPU development, goes to datacenters made to house them for the purposes of training and inference
These are secondaries that are driven indirectly by the "AI demand", and it is the existence of the latter that isn't obvious to me. The primary driver is the
and the products built off of them,
which are, IMHO, few and surprisingly underwhelming.
include medical research, physics research, customer service, and development
I'm seeing the "physics research" first hand. Most of it is complete bullshit, especially the arm that has given up on experiment design and instead has switched to continuous noise recording which is then massaged with "AI" to find "signals". Often confirmed by the night shift on the same beam that f
Re: (Score:2)
These are secondaries that are driven indirectly by the "AI demand", and it is the existence of the latter that isn't obvious to me. The primary driver is the
Indeed, they are secondaries driven by AI demand. But they're still part of the final accounting.
which are, IMHO, few and surprisingly underwhelming.
I find that incredible.
For better, or for worse, they've changed the direction of technology, and even science.
I'm seeing the "physics research" first hand.
Oh, ya?
Most of it is complete bullshit, especially the arm that has given up on experiment design and instead has switched to continuous noise recording which is then massaged with "AI" to find "signals".
You've just described most of particle physics and astronomy, and you're about 4 decades late in your complaint.
Discoveries have been made by pumping past collected data that nobody had any idea how to process into full into neural networks to tease out the signal.
Often confirmed by the night shift on the same beam that feeds on the same budget.
The "night shift" ain't confir
Re: (Score:2)
You've just described most of particle physics and astronomy
What else is out there?
, and you're about 4 decades late in your complaint.
Incidentally, the very 4 decades that have brought only more deferents and epicycles.
The "night shift" ain't confirming shit out of things like ATLAS, not unless the night shift is about 150,000 people.
The "night shift" of ATLAS is CMS. Sometimes it is even ahead and shows ATLAS which data to massage.
What's the serious value proposition in Social Media?
Control. Haven't you noticed how its billionaire owners took over your country completely in 3 short months?
The value proposition for LLMs (which I think you're attacking here) are that kids are using them to augment their skills (and replace, sadly)
Hard to argue that isn't valuable- to them.
It isn't. They'll figure it out eventually.
The market is exploding now.
The investment bubble is exploding now, which ain't quite the same thing.
Re: (Score:2)
What else is out there?
I think that was my point...
Incidentally, the very 4 decades that have brought only more deferents and epicycles.
Wrong. If you had held back the "only", I'd have let you have that one, because ya, we've gotten lots of epicycles.
The "night shift" of ATLAS is CMS. Sometimes it is even ahead and shows ATLAS which data to massage.
And both use ML!
Do you think CMS has guys pouring over that data?
The damn thing produces 500TB of data per second.
Control. Haven't you noticed how its billionaire owners took over your country completely in 3 short months?
No, from the perspective of those that allow themselves to be monetized by it.
We don't care what the value proposition for the owners are. They can't make any money without the sheep. The value is in the sheep.
It isn't. They'll figure it out eventually.
Just like they've figured out that being m
Diplomatic class seeking status and attention (Score:2)
It's another example of the UN, NGO, nonprofit and the diplomatic class seeking status and attention.
The UN's claims from the article
- AI may displace a lot of workers
- AI spending and expected economic productivity benefits is concentrated in a few countries
Simply, the UN's press release can be restated as the old mantra "The money is concentrated in a few countries and it is unfair that people in countries who did no work to earn it, deserve via charity an unearned wealth transfer. The UN is willing to v
Has AI ever done anything appreciable... (Score:3, Insightful)
...without an external validator to ensure accuracy? I have never seen AI perform any activity beyond trivial examples in the training canon in a self validated manner. This is an honest question, can anyone present a reliable novel ability of an AI agent?
Just for clarity, I am not asking about a cherry picked one-off it did this one thing this one time... I want examples of a reliable novel ability.
Re: (Score:2)
be capital, n00b! (Score:2)
Obviously the rich will continue to absorb the wealth from those productivity gains, but at some point something's got to give.
This has been going strong 65 years [epi.org], why change a winning strategy?
"A body remains at rest, or in motion at a constant speed in a straight line, unless it is acted upon by a force." [wikipedia.org] I'll let you draw what inference you may from the highlighted clause. The majority of Americans are too dense to do so.
Re: (Score:2)
If it can do all the basics of your desk job, the chores that you spend most of your time on that don't really require thought, just action, you'll either be working two hours per day or replacing four workers like yourself.
The problem that no one selling a product will ever acknowledge is that AI can't do any of those things reliably enough to actually save anyone time. You either end up wasting at least as much time baby sitting or even more time cleaning up the mess.
"Oh, but it will get better!" Not with current approaches. There are real limits there. Neither is progress inevitable, as the doe-eyed faithful love to claim. AI companies are hemorrhaging money at the moment, the hope being that they'll figure things out
Re: (Score:2)
I thought I made it pretty fucking clear that you don't get to act like you're an expert anymore.
"Oh, but it will get better!" Not with current approaches.
All available evidence to the contrary.
There are real limits there.
Of course there are. But you don't know what they are.
Neither is progress inevitable, as the doe-eyed faithful love to claim.
There are only a few things inevitable in this world- death, taxes, and you pretending like you're an expert in this topic, in spite of a quite humorous collection of past commentary demonstrating the exact opposite.
Time to go poof again.
Re: (Score:2)
When some asshole who goes around calling everyone who doesn't take his word as a self-proclaimed published expert on the topic a moron, dumbshit, doe-eyed faithful, and a litany of other putdowns, can be shown to be a bullshit artist by merely referencing his historical statements on the topic which directly contradict themselves more times than one can count, including directly stating things they have called people a moron for believing, you shut that fucker down.
You don't
Re: (Score:3)
You're mixing up usefulness and innovation. AI isn't creating anything on its own, but it's making the process of creation faster. It's just a tool that makes people a little more efficient at what they do. If I have a task to do, such as writing code, and AI can spit out boilerplate code for me that I can just build on, then that made me more efficient at my work.
Yes, sometimes it spits out garbage, but with time and experience one learns how to wrangle it and get it to spit out useful bits, and that's pre
Re: (Score:3)
AlphaFold2 is the most obvious example. It was validated to work properly, but the required validation work is nothing compared to the work it saved by predicting folding of hundreds of millions proteins, which each take about half a year of work. If you are not familiar of AlphaFold2, I suggest finding a video or story about it, because I think it is much bigger thing than landing to the Moon was.
If you are talking about (multimodal) LLM. The new Gemini 2.5 is something that can actually be used to create
Re: (Score:2)
What do you think how the text to speech system the hotlines use work? There are also many use-cases for OCR that are not more validated than using a simple spellchecker to detect misrecognized words (which are rare with AI based OCR). Many segmentation tasks don't need correction, but the interesting ones usually nevertheless have control, because many of these tasks are too important to not have a human check it.
Re: (Score:2)
I do software development with AI assistance every day now and it frequently works quite well. It often saves me hours or even days. I've seen very large improvements in the capabilities of AI just over the past few months. An enormous difference over the past year.
AI is "computers" (Score:1)
50+ years ago they said computers would replace all our jobs. Also, robots, will replace us. And FSD, drones, crypto, automated checkouts, online music, Japan, China, immigrants, Communists, Jews, fast food, no God in schools, books describing sex acts, Swifties, etc; They are all threats and we should be scared. At least, that's what we're being told by the "experts".
Re: (Score:3)
And people still have work. Some even to keep the computers running.
Re: (Score:1)
Speak for yourself. Most of the people I know (and I was working for a F500 company, and the only reason I have a job is because of who knows me) are out of work. I know people who were making $300k/year at the start of 2024 doing mall Santas come Christmas.
Jobs are an exception, not the rule. Yes, it is fun to make fun of the schmuck who blew his brains out because he couldn't afford the payments on his BMW, but this is getting to be past 2020 with so many people out of work, maybe even heading for 1980
Re: (Score:2)
You've moved the goalpost. You're not talking about the jobs the robots made obsolete in this post. ;)
Even though I think in Futurama there were a robot santa I am not sure if it were in the mall.
Re: (Score:2)
"Yes, it is fun to make fun of the schmuck who blew his brains out because he couldn't afford the payments on his BMW"
Is it?
Re: (Score:2)
Computers and robots have replaced a lot of workers.
1. Have a look at Amazon. They have a lot of robots there. Those robots are replacing robots https://www.youtube.com/watch?... [youtube.com]
2. Amazon has also a website, that is replacing a lot of small businesses with the help of those warehouse robots.
3. China has replaced a lot of factory workers with robots
4. Data input and processing that was previously done by humans, is not handled by computers is many locations.
5. Automated sorting machines in the post office
6.
Widen Inequality? (Score:1)
Yeah, maybe (Score:4, Informative)
Predictions are hard, especially about the future, and it's getting increasingly unpredictable
If monopolists and governments control the tech, prices will be high and only the rich will benefit
If open source continues to gain acceptance, and the tech is available to all, all will benefit
Re: (Score:1)
Inequality, yes - but not as you know it, Jim (Score:2)
The thing about LLM's is that they are teriffic if you know things already, being just mediocre at something will make you fantastic at your work if you know how to use it.
But it will create a very large shift in equality, for example:
- People with zero skills and afraid of A.i. and LLM's will miss out, they will have a hard time competing with people that has just a mild interest in using it as a tool.
- You will see a huge amount of "dummies" that will think this thing is alive, and they'll have a new best
Re: (Score:2)
Those that are already good at something, stand a chance of becoming 10-100 times better than they were
That seems unlikely. It turns out that if you're competent enough to vet the output, you'll quickly discover that you're wasting more time than you're saving trying to get the things to spit out anything useful.
I can't remember where I saw this, but it was something like "The most amazing thing about AI is while it's obviously an expert in things I know nothing about, it obviously knows nothing about things I'm an expert in." It puts me in mind of Gell-Mann Amnesia [wikipedia.org].
Re: (Score:2)
Could very well be.
But one thing I do know, is that I am in NO way more competent than the 1000s of books of programming tips and knowledge there is out there, and that no matter how little or much I think I know, I learn all the time, and sometimes it can fast-track me towards what I was seeking, and suggest things I wasn't even remotely aware of.
Competence is many things, also the ability to not overevaluate yourself.
Re: (Score:2)
Along with things that don't exist and things that are completely wrong. As a search engine, it doesn't get much worse.
Re: Inequality, yes - but not as you know it, Jim (Score:2)
Re: (Score:2)
All technological advances have always led to inequality, but also have made lives better for the majority of people. The industrial factory made life worse for the blacksmiths, who were put out of business, but it made life better for the rest of humanity, who could get horseshoes and nails much more cheaply than before. Even the blacksmiths themselves benefitted from this lower cost.
Dispirit impact is unavoidable with any business or any technology. That should not lead us to shun technology, but rather,