OpenAI CEO Tells Federal Reserve Confab That Entire Job Categories Will Disappear Due To AI (theguardian.com) 70
An anonymous reader quotes a report from The Guardian: During his latest trip to Washington, OpenAI's chief executive, Sam Altman, painted a sweeping vision of an AI-dominated future in which entire job categories disappear, presidents follow ChatGPT's recommendations and hostile nations wield artificial intelligence as a weapon of mass destruction, all while positioning his company as the indispensable architect of humanity's technological destiny. Speaking at the Capital Framework for Large Banks conference at the Federal Reserve board of governors, Altman told the crowd that certain job categories would be completely eliminated by AI advancement. "Some areas, again, I think just like totally, totally gone," he said, singling out customer support roles. "That's a category where I just say, you know what, when you call customer support, you're on target and AI, and that's fine." The OpenAI founder described the transformation of customer service as already complete, telling the Federal Reserve vice-chair for supervision, Michelle Bowman: "Now you call one of these things and AI answers. It's like a super-smart, capable person. There's no phone tree, there's no transfers. It can do everything that any customer support agent at that company could do. It does not make mistakes. It's very quick. You call once, the thing just happens, it's done."
The OpenAI founder then turned to healthcare, making the suggestion that AI's diagnostic capabilities had surpassed human doctors, but wouldn't go so far as to accept the superior performer as the sole purveyor of healthcare. "ChatGPT today, by the way, most of the time, can give you better -- it's like, a better diagnostician than most doctors in the world," he said. "Yet people still go to doctors, and I am not, like, maybe I'm a dinosaur here, but I really do not want to, like, entrust my medical fate to ChatGPT with no human doctor in the loop." [...] At the fireside chat, he said one of his biggest worries was over AI's rapidly advancing destructive capabilities, with one scenario that kept him up at night being a hostile nation using these weapons to attack the US financial system. And despite being in awe of advances in voice cloning, Altman warned the crowd about how that same benefit could enable sophisticated fraud and identity theft, considering that "there are still some financial institutions that will accept the voiceprint as authentication".
The OpenAI founder then turned to healthcare, making the suggestion that AI's diagnostic capabilities had surpassed human doctors, but wouldn't go so far as to accept the superior performer as the sole purveyor of healthcare. "ChatGPT today, by the way, most of the time, can give you better -- it's like, a better diagnostician than most doctors in the world," he said. "Yet people still go to doctors, and I am not, like, maybe I'm a dinosaur here, but I really do not want to, like, entrust my medical fate to ChatGPT with no human doctor in the loop." [...] At the fireside chat, he said one of his biggest worries was over AI's rapidly advancing destructive capabilities, with one scenario that kept him up at night being a hostile nation using these weapons to attack the US financial system. And despite being in awe of advances in voice cloning, Altman warned the crowd about how that same benefit could enable sophisticated fraud and identity theft, considering that "there are still some financial institutions that will accept the voiceprint as authentication".
Final Harvest (Score:5, Funny)
"The OpenAI founder described the transformation of customer service as already complete".
In a similar way, I might describe the transformation of last night's dinner as already complete.
Re: (Score:2)
Yeah, I'd imagine that they'll soon be a severe backlash to AI customer support bots.
I think that some businesses will keep advertising that their customer support is staffed by "Real humans in North America" because they know that they're clients are sick of AI call trees and outsourced customer support that can barely speak English.
Re:Final Harvest (Score:4, Interesting)
Re: (Score:1)
The backlash doesn't matter too much to them in the same way that Google doesn't care about the fact that your problem needs a human. They have a release valve in the forums, where they can see what your problem is, measure it and decide that, since it only affects 0.1% of the customers, they don't care.
The only reasonable way to fight this is through the courts. Specifically both normal and small claims / cheap courts where it's possible to actually force companies to do what needs to be done. That's where
Re: (Score:2)
Indeed. Just yesterday the great Amazon "customer service" chat AI failed to understand (by its own direct admission) "article xyz was not in package" and had to hand me off to a human.
If that is the transformation that is "complete", I am not impressed.
Serious (Score:2)
Is anyone taking this seriously? If so, we are experiencing the biggest scam of the modern age (whatever that hyperbole means.) What will the historians call this? The utterly fucked up 20s? The scam to beat all scams? Or will our future historians just say, "thank god we are not that stupid anymore."
Re: (Score:3)
Re: (Score:3)
Sam Altman has the world's most expensive barrel of snake-oil and he's determined to sell all of it. ChatGPT and other LLMs like it have some very, very niche cases where they can effectively replace human workers. Remote customer service where the role is very simple and well-defined, employees essentially act programmatically following a rigid procedure, using "business friendly" language... that's obviously the one that maps most closely to an LLM's core competency.
I'm absolutely sure there have been "ex
Re: (Score:2)
They weren't AI-based, they were just programs where you answered a bunch of questions about the patient, symptoms, test results etc. and it came up with the most likely diagnoses.
I hate the break it to you, but expert systems are AI. In fact, expert systems were the driving force behind the great AI Boom 40+ years ago. See: XCON [numberanalytics.com]
They're expensive to make and expensive to maintain, but expert systems don't hallucinate. They're utility and reliability are why they're still in use, with new systems being developed and deployed all the time. They don't get a lot of attention in the press, but they are still and active area of research.
Classic scam (Score:3)
>> The question here is, are they CHEAPER than the currently option? That costing is complicated, and it really starts on how the AI companies price each query.
That's the classic silicon valley scam:
- Burn VC money to undercut the existing options.
- Raise prices after everyone has switched over and the other options are gone.
Re: (Score:2)
"done correctly" is doing a LOT of heavy lifting here.
Re: (Score:3)
What will the historians call this? The utterly fucked up 20s? The scam to beat all scams? Or will our future historians just say, "thank god we are not that stupid anymore."
Futurama labeled it "the stupid ages". I think that's about right.
Re: (Score:2)
What will the historians call this? The utterly fucked up 20s? The scam to beat all scams? Or will our future historians just say, "thank god we are not that stupid anymore."
Futurama labeled it "the stupid ages". I think that's about right.
I think that fits very well.
3rd industrial revolution (Score:5, Insightful)
If you look at the through line between the first two industrial revolutions you're going to find massive technological unemployment that we all just gloss over in the history books. You don't even begin to get a discussion of it until 200 level history courses.
I've forgotten most of it but I briefly dabbled with being a historian in college so I took a few of those courses.
And yeah we have massive amounts of technological unemployment until the two world wars killed enough people and blew up enough stuff that we got back to full employment.
At the time we only had about 3 billion people so there was quite a bit of land to expand into. We've used up that land. There is still plenty of space but not the super cheap readily available resources that the greatest generation and boomers were able to expand into in order to maintain full employment.
And we certainly do not have the government programs and spending needed to maintain that expansion even if we had the space. Because that's communisms.
So yeah we are about to have another round of massive technological unemployment and that's going to create a massive amount of social unrest. On top of that climate change is causing droughts which will put pressure on the food supply.
You can already see the cracks forming with democracy's collapsing around the world. Wars are starting back up with Russia doing the first major land grab war in decades. And the rest of the world just kind of let them do it. Because we're busy with our own fascists.
Eventually the social unrest will bubble over and fascists will get put in charge because they will promise simple solutions to complex problems.
That will result in an economic collapse because fascists are as bad at running countries and they will conduct wars to loot other nations in order to fill their coffers and keep their citizens under control.
Only this time we've got nukes.
The fact that folks don't know history and haven't bothered to look into what llms can actually do isn't going to stop any of the above from happening. No matter how far you put your head in the sand it's still going to get lopped off
Re: (Score:1)
1. You need to physically be there to do it
2. You need a reasonably difficult qualification most don't want to study
3. You need a license to practice governed by a small board of corrupt fellows out to maximise their incomes
Re: (Score:2)
Re: (Score:2)
He's in the business of AI ... (Score:2)
Sure...yea...that's the actual experience. (Score:2)
Altman told the crowd that certain job categories would be completely eliminated by AI advancement. "Some areas, again, I think just like totally, totally gone," he said, singling out customer support roles. "That's a category where I just say, you know what, when you call customer support, you're on target and AI, and that's fine." The OpenAI founder described the transformation of customer service as already complete, telling the Federal Reserve vice-chair for supervision, Michelle Bowman: "Now you call o
Re: (Score:2)
Re: (Score:2)
That's fair for mandatory spending. For discretionary spending I think capitalism will weed out those unwilling to support their customers.
So sure, the single choice internet provider serving you can do that, but the place selling you jeans can't.
Re: (Score:2)
Lol. The market won't save you. Nobody will move for better customer service if the total price is higher. And come on, how often do you call a customer service line about jeans? It's telcos, energy, insurance, etc... they're the ones that'll use this, and once they're all using it, it absolutely won't matter that customer satisfaction drops off a cliff as soon as the reason for the call falls outside the normal parameters.
Re:Sure...yea...that's the actual experience. (Score:5, Insightful)
There are plenty of discretionary services that require support. Do I need a phone? No. Do I need cable TV? No. Streaming services? Nope. Amazon? Nope. Can I shop in stores instead of online because the online CS sucks? Yes.
People absolutely will pay more for better service. There is a reason I bought a BMW. It was the service center in town. The Audi service center treated me like a replacable comodity. I switched housekeepers to one more expensive simply because the previous one was rude. I switched landscapers because I didn't want a truck with a politcal flag swinging around my yard.
I shop at Ace hardware over the big box stores because they have staff that answers questions.
Service might not matter to you, but it does matter to many of us.
Re: (Score:2)
I shop at Ace hardware over the big box stores because they have staff that answers questions.
That's what we do, and we're very successful at it. Every time the Orange Suck up the street runs a TV ad, we see a bump in sales as a direct result.
Re: (Score:2)
Re: (Score:2)
How so? I haven't had a phone call I've answered in years. A phone is useless. An internet connection is useful.
So you're assuming something you shouldn't (Score:2)
So yeah you can shop at Ace hardware but for how much longer? The reason you can find somebody to talk to there is because their prices are a bit higher, which is nice that you can afford that, and the private owner probably hasn't seen a buyout yet.
Eventually that old guy whose owns your local Ace is g
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Then there is no hope (Score:2)
Altman is telling us that leaders will merrily do what ChatGPT says to do, just as hostile nations tell them to do
His vision of the future includes a feature that everyone using ChatGPT will self destruct.
Re: (Score:2)
Re: (Score:2)
Well, given the brilliant Senator from Utah just fell for a AI generated "resignation letter" from the Fed chairman Powell, I don't think it takes much for a hostile nation to trick our political class. Seriously, Mike Lee, you might want to check facts before you open your trap.
I wonder if Sammy told the good senator that his job is replaceable by AI?
Re: (Score:2)
The good news is, this statement is nothing more than a CEO claiming that his company's product is great. As with all marketing claims, reality is significantly less rosy (or in this case, dreadful) than the marketers would have you believe.
Re: (Score:2)
The good news is, this statement is nothing more than a CEO claiming that his company's product is great. As with all marketing claims, reality is significantly less rosy (or in this case, dreadful) than the marketers would have you believe.
What I don't get is why something that is so obviously a fatal flaw to me is spouted out and listened to without a problem. Altman should be glad I wasn't there. 8^)
Re:Then there is no hope - proves (Score:2)
what does AI do when... (Score:2)
What does AI do when a customer calls up and describes the symptoms of some techno-problem, like their router to the fiber link to their house is screwing up. The symptoms do not scream, "yo, you router is kapitski". Rather, does it (1) run through all the inane scenarios it knows about which it has a chance of fixing and then start at the top when you cannot give it that blessed symptom that screams "yo, you router is kapitski", or (2) finally give up and tell you to call a human representative, or (3) act
Re: (Score:2)
They are going to fire the human representatives
Re: (Score:2)
They are going to fire the human representatives
Maybe they'll add the AI generated call center transfer hell. You just keep getting passed to another AI agent that says you're talking to the wrong department.
Re: (Score:2)
I would expect the AI to say "fix the core routing network" based on the last problem I had. It turns out the ISP's config is broken for IPv6 BGP via two different backends on their system to the same router on my end. I kept getting the BGP packets on the wrong interface so one link would never come up. I have no idea how that would happen but it did. Oddly the v4 BGP works quite well.
Re: (Score:1)
" I have no idea how that would happen but it did."
You should ask an AI how it happened.
It does not mistakes (Score:5, Insightful)
The OpenAI founder described the transformation of customer service as already complete, telling the Federal Reserve vice-chair for supervision, Michelle Bowman: "Now you call one of these things and AI answers. It's like a super-smart, capable person. There's no phone tree, there's no transfers. It can do everything that any customer support agent at that company could do. It does not make mistakes. It's very quick. You call once, the thing just happens, it's done."
Yea, AI would never make up references, cheat, hallucinate or quote ridiculous prices... It's perfect.
Re: (Score:3)
Re: (Score:3)
I think we have moved from AI's hallucinating to AI founders hallucinating.
You have your timeline reversed. AI founds were hallucinating before they founded their companies.
Doesn't make mistakes? (Score:2)
We've already seen companies get in trouble for the mistakes of AI chatbots...
https://www.cbsnews.com/news/a... [cbsnews.com]
In fact, I'm sure there is going to be a cat and mouse game of AI chatbot hacking for customer service. People could use chatbots to contact customer service chatbots and manipulate them into giving discounts or other unauthorized things. Likewise, we know there will be plenty of frustrated customers who can't get the chatbot to display basic understanding a human would have.
Re: (Score:2)
Re: (Score:2)
Re: (Score:3)
Tungsten cubes [medium.com], that's what can go wrong.
This man is clearly psychotic (Score:2, Interesting)
That won't happen, of course; as s society we've already decided that it's perfectly fine to have psychotics, sociopaths, megalomaniacs, and morons (e.g. RFK Jr, Musk, Zuckerberg, Trump) in positions of leadership. And we'll go through today and tomorrow and the next day pretending that this is normal and okay while they drive not just IT, not just the US, but the entirety of human society over a cliff.
Nope (Score:1)
But does ceo speak with forked tongue? (Score:2)
Support is often due to lack of functionality (Score:2)
"It does not make mistakes" (Score:2)
What world are these people living in? I use ChatGPT every day and it is ALWAYS making mistakes.
Re: (Score:1)
Are you using the free version?
I have a suite of "pro" LLMs at work, they are notably smarter than freebies. But they do a lot of mistakes too though.
Pretty much anything that was not previously discussed on the internet is likely to lead nowhere.
Re: (Score:2)
Yes I pay for ChatGPT pro. Like I said, I use it every day....mostly for my coding job. I do find it incredibly useful. But you have you really can't trust the output. You have to make sure to always check the code and source the output.
CEO of AI company says his product is awesome (Score:2)
What else would we expect?
Reality is significantly less dramatic than CEO marketing claims.
The federal reserve needs a serious haircut? (Score:1)
As of the most recent data from 2023 and early 2025, the Federal Reserve System employs approximately 23,900 to 24,000 people across its entire system.
- Perplexity
They regulate banks and the money supply. Do they really need 24k people to do that? Is the fed still needed and at this size?
BTW, their regulation of banks, specifically online bank security sucks. No US financial institution should still be using SMS or phone calls for 2FA.
Sure (Score:2)
I'm pretty sure about that. But some of these categories were also not existing, e.g., before the Internet. Some jobs come, some jobs go. In the past you needed more scribes and fewer mail server admins.
Lots of Potential Yet Still Over-hyped (Score:2)
Contemporary AI, and particularly generative AI, has extraordinary potential but can only fully do a job in a small number of roles. Those roles are where it is used as a tool.
What I mean by a tool is that it is driven by assigned goals rather than free will value judgments. This makes it useful as an assistant to a human, at best. An additional limitation is its inability at fundamental creativity/innovation.
Intelligence is that ability to solve problems. A human mind is centered around something more
First to go ... (Score:1)
Whole categories (Score:2)
Will one of the categories eliminated be the dipshit prognosticator CEO who is constantly touting that his product will lead to massive societal revolution? I'd be all for that.
If not, I'd also accept removing the category of journalist who writes articles about the CEO's self-serving drivel. It seems like AI would be perfect for either of those positions.
No Doubt (Score:2)
That a lot of companies will try. And LLM could, probably, replace some barebones-type customer service ("have you tried to turn it off and then back on again?").
I guess it depends on a lot of things, like how complex the issue or situation is. Or what kind of permissions are granted to the LLM (would it be allowed to order parts for you? Could it send certain parts free? Could it issue credits?). On top of that is liability. What happens with the LLM does something outside the normal process? People are f