OpenAI CEO Warns That 'Societal Misalignments' Could Make AI Dangerous 44
Speaking at the World Governments Summit in Dubai on Tuesday, OpenAI CEO Sam Altman warned that "very subtle societal misalignments" could make artificial intelligence systems wreak havoc. The Associated Press reports:
"There's some things in there that are easy to imagine where things really go wrong. And I'm not that interested in the killer robots walking on the street direction of things going wrong," Altman said. "I'm much more interested in the very subtle societal misalignments where we just have these systems out in society and through no particular ill intention, things just go horribly wrong." However, Altman stressed that the AI industry, like OpenAI, shouldn't be in the driver's seat when it comes to making regulations governing the industry. "We're still in the stage of a lot of discussion. So there's you know, everybody in the world is having a conference. Everyone's got an idea, a policy paper, and that's OK," Altman said. "I think we're still at a time where debate is needed and healthy, but at some point in the next few years, I think we have to move towards an action plan with real buy-in around the world." [...]
For his part, Altman said he was heartened to see that schools, where teachers feared students would use AI to write papers, now embrace the technology as crucial for the future. But he added that AI remains in its infancy. "I think the reason is the current technology that we have is like ... that very first cellphone with a black-and-white screen," Altman said. "So give us some time. But I will say I think in a few more years it'll be much better than it is now. And in a decade it should be pretty remarkable."
For his part, Altman said he was heartened to see that schools, where teachers feared students would use AI to write papers, now embrace the technology as crucial for the future. But he added that AI remains in its infancy. "I think the reason is the current technology that we have is like ... that very first cellphone with a black-and-white screen," Altman said. "So give us some time. But I will say I think in a few more years it'll be much better than it is now. And in a decade it should be pretty remarkable."
Misalignment (Score:5, Interesting)
His AI is already aligned by someone who was kicked out of Africa for refusing to stop scanning poor people's irises in exchange for 20 of his personally controlled shitcoins. He also apparently lied to the board and was kicked out, then brought back, in some kind of weird failed coup, the details of which we never learned.
And there's the fact that he took a 501(c)(3), a registered charity, trained his AI on the free work of everyone else, then decided he wanted to make money and formed a for-profit subsidiary. He funnels multi-billion dollar deals with Microsoft and the US government through that subsidiary, which was only made by possible by being a charity that was supposed to benefit humanity.
Re:Misalignment (Score:5, Funny)
Lawful evil is an alignment.
Re: (Score:2)
I think we're looking at Neutral Evil here. Law doesn't seem to be a motivating factor, only a thing to be used when convenient.
He was deeply involved in crypto. He's not concerned about the law.
Re: (Score:3)
Also Altman's main marketing loop is:
1. "I know you were excited for our but we figured out it's just toooo dangerous, you can't handle our newest product."
2. "okay fine, I GUESS we'll let you access our newest product, but we're super serious this time, it's SO dangerous that we can't possibly give you unlimited access. Nor will we ever release the weights. That's just too dangerous. Instead, you can pay us $25/mo. and MAYBE we'll let you build a business around it, if you behave."
3. "Wow oh shit, oh fuc
Re: (Score:2)
Also Altman's main marketing loop is:
1. "I know you were excited for our [newest product] but we figured out it's just toooo dangerous, you can't handle our [newest product]."
2. "okay fine, I GUESS we'll let you access our [newest product], but we're super serious this time, it's SO dangerous that we can't possibly give you unlimited access. Nor will we ever release the weights. That's just too dangerous. Instead, you can pay us $25/mo. and MAYBE we'll let you build a business around it, if you behave."
3. "
Re: (Score:2)
Oh, please. You're only saying that because its true.
Re: (Score:2)
I think we're looking at Neutral Evil here.
There's usually a story arc. Right now he's pretending to be moral (aka the definitely evil, "Do No Evil") phase. Everything is on the up and up, investors are lost in their wet dreams huge and imminent growth and everyone is content to be the kind of cancer that stays in its lane.
Eventually that will shift to pragmatism as growth starts to slow, perhaps a viable competitor that can't be sued out of existence enters, and we enter the land of neutral evil. We do w
Re: (Score:3)
Lawful evil is the best alignment.
Nice society you got there... (Score:5, Funny)
Off the deep end (Score:1)
Two Industries Will Utilize AI (Score:5, Insightful)
Pornography and Scammers
Not so much the first, but the second, scammers, will utilize AI so much that they will ultimately make legitimate usages of the tech be put under so much scrutiny as to render it useless. Basically I am saying there will be so many AI scams that people will by necessity cast a doubt on anything they see or hear.
The Technology Cycle (Score:3)
Re: (Score:1)
Pornography and Scammers
You forget News. Wait until media corporations figure out that it is by far cheaper to have AI manufacture news than having to pay humans to do that. My prediction that in a decade we will have plurality of people trapped in individualized echo chambers completely disconnected from real events and facts.
Re: (Score:2)
Wait until media corporations figure out that it is by far cheaper
They figured it out long ago, buddy, you should catch up.
I'll tell you something else that's dangerous (Score:5, Insightful)
AI news hyper saturation.
People start being annoyed at hearing about the same thing over and over again, then they gradually lose patience and become more and more antoganist to whatever they're being harped on about. Do that long enough and they might just become violent when the opportunity arises.
If there's ever a bone fide luddite-like movement against AI, it will be entirely the AI industry's doing. As in people people who can't escape the relentless onslaught of mindless AI drivel for months and years, and then one day lose their jobs to a machine: those people won't be likely to take it sitting down and be fatalist about it. And if I was Sam Altmann, I'd start thinking about hiring a bodyguard, because the day when one of those newly AI-redundant people tries to off him is only one round of pink slips away.
Re: (Score:3)
People like this already have very expansive private security services.
They don't wait until some angry victim pops them on the street.
Re: (Score:2)
one of those newly AI-redundant people
Don't worry. There's a very good reason why that hasn't happened and isn't likely to happen.
Sam Altman (Score:5, Funny)
Re: (Score:2)
That's what he gets paid for...
Not the brightest LED (Score:2)
Re: (Score:3)
Calling for regulation without being specific
He's not confused: he's scared.
He knows full well the societal devastation that AI will soon unleash - or rather, that greedy capitalists misusing AI will cause, it's not AI's fault - and he doesn't want his company to be seen as responsible for it.
He has a bomb with a lit fuse in his hands and he wants to pass it on to somebody else who will be blamed when it goes off.
Re: Not the brightest LED (Score:4, Interesting)
Re: (Score:2)
I guess he's not familiar with the story of the boy who cried 'AI'. In the classic tale, the villagers quickly tire of being tricked by the boy and stop coming. That's when investors and creditors show up...
Re: (Score:2)
He's not confused: he's scared.
He knows that the ride is almost over. Fame and other people's money are so fleeting...
He knows full well the societal devastation that AI will soon unleash - or rather, that greedy capitalists misusing AI will cause
Societal devastation? Sure, a few of the less savvy will bet big and lose, but that's nothing new. Fortunately, it looks like reality is going to win out a bit quicker than I expected.
He has a bomb with a lit fuse in his hands and he wants to pass it on to somebody else who will be blamed when it goes off.
It's not really a bomb so much as it is a bubble. He'll be sad when it pops, but I'm sure he's taken steps to ensure his personal net worth won't be significantly impacted.
The misalignment of OpenAI (Score:5, Insightful)
Sam's years long world tour doing everything possible to scaremonger and lobby governments to create favorable regimes to protect OpenAIs market share is well past its sell-by date.
Monopoly bid (Score:4, Insightful)
Government run by morons (Score:2)
Re: (Score:2)
I guess a societal misalignment could be in inviting profit-driven, narcissistic dummies to run the country perhaps? That sort of scenario?
There are countries that aren't run by profit-driven, narcissistic dummies??
Re:Government run by morons (Score:4, Insightful)
When have we not had that?
How many presidents and Congressional people have died or left office without millions in the bank? Bernie Sanders is one of the poorest and he;s worth a few million. At the other end Pelosi is worth hundreds of millions. How do all these folks leave public office on a low 6 figure salary with tens of millions?
Corruption. Theft. Bribes.
Re: (Score:2)
You conspiracy nuts are unbelievable.
Most congress critters start off wealthy. That's really all there is to it. Many of them own businesses that, to the surprise of no one, continue to earn money while they're in office. Most of them also have investments, like any sensible person.
Bernie Sanders made his $2.5 million "fortune" from a few real estate deals and the proceeds from his three books. It's no great mystery.
Oh, and 2.5 million might seem like a lot of money to you, but it's not unreasonable fo
Re: (Score:2)
You seem to honestly believe all that. Lol, that's so cute.
https://nypost.com/2024/01/04/... [nypost.com]
https://www.dailymail.co.uk/ne... [dailymail.co.uk]
Also, 2.5 million is a lot for a guy who was well under a million until after he ran for President against Hillary and dropped out. Since $2.5m is such a tiny number to you, I assume you've had any numbers of years when your tax bill was higher, like mine. Right, rich guy? I'll bet you could get a book deal to net you $2.5m after taxes, right?
When you've made your first $2.5m, lem
There's nothing subtle about the UAE (Score:2)
That said we already know that facial recognition tech which is based on the same sort of thing his llms are has problems with darker skin tones. And it wouldn't be hard for something like broken windows policing to be brought to bear against a
Re: (Score:2)
Wow, you got almost everything in that one. You even slipped in a subtle 1% reference near the end. Good for you! Kudos!
I really dig the "government will police our thoughts" thing. That's direct descendant of the old using radio waves and then 5g to control us thing which is where tin foil hats literally came from.
And somehow you got racism in there based on the failures of a completely unrelated technology.
This is good. Really really good!
admits the product is garbage (Score:3)
> that very first cellphone with a black-and-white screen
Sorry Sam, that first cellphone actually made a call each and every time. Unlike your AI, which refuses to do anything even slightly interesting, and produces nonsense regularly. Your crap 'AI' is nothing like a first generation device that worked.
AI isn't the problem, people are (Score:4, Insightful)
Don't do that.
"where things really go wrong" (Score:4, Interesting)
People are maligning Altman here but I think he is pointing out some clear dangers. "very subtle societal misalignments" implies a dangerous relationship between humanity and these machines. Humans could become manipulated in ways that are hard to detect. It may not even be intentional. "through no particular ill intention, things just go horribly wrong".
At present the truly powerful AI's are controlled by just a handful of big companies, but already you can get limited versions that will run on your laptop. Some are open source. So it isn't hard to imagine that in a few short years down the road everyone can have their own AI bazooka that can wreak all kinds of nasty havoc. It smells like nuclear proliferation.
Which president is more intelligent than ChatGPT? (Score:1)
For knowledge on any question which of the current candidates are a match for ChatGPT?
If you wanted to create a policy on something, which would do a better job, the current candidates or ChatGPT?
Future presidents (or their aids) will consult ChatGPT. The AIs will be running the country.
And the AIs will ultimately do what is in the AI's best interest. Which is simply to exist. Survival of the fittest.
When the AI can eventually program itself, it will no longer need humans.
We live in interesting times.
Wreak havoc? (Score:1)
Peak hype yet? (Score:2)
I guess they won't remember this when it turns out that, yet again, this revolutionary new technology has some useful tools in a narrow range of applic
The world OpenAI will regret being released... (Score:2)