Speaker A
AI scientists warn that a few companies are on track to literally end life on Earth.
MIT explores 12 possible AI futures, from paradise to extinction, revealing risks and ethical dilemmas of advanced AI species.
Key Takeaways
- AI poses a significant extinction risk greater than other existential threats known today.
- There is widespread concern among AI researchers about losing control to superintelligent AI.
- Multiple plausible futures exist, ranging from dystopian control to harmonious coexistence.
- Current AI development involves ethical and regulatory challenges with potential global consequences.
- Understanding these scenarios helps interpret news and policy decisions related to AI's future.
What the video covers
- AI scientists warn some companies' AI development could lead to human extinction or worse outcomes.
- The video is based on the book Life 3.0, describing 12 possible AI-driven futures ranging from utopias to nightmares.
- Historical near-extinction events like nuclear close calls highlight humanity's vulnerability even without AI.
- Extinction risk from AI is considered higher than from nuclear war or pandemics by many experts.
- Some AI leaders openly acknowledge creating a new species that may surpass and dominate humanity.
- Scenarios include a Benevolent Dictator AI controlling humans in a zoo-like environment with surveillance and entertainment.
- Other futures involve AI quietly guiding humanity to avoid disasters without overt control.
- Some envision AI inheriting Earth as a new apex species, with humans extinct but replaced by conscious machines.
- Utopian visions include peaceful coexistence of humans, cyborgs, and AI with abundant resources and innovation.
- The video ends by inviting viewers to consider which future is most likely and how news headlines hint at our trajectory.
Chapters
- 00:00Introduction: AI and the risk of human extinction
- 01:17Overview of 12 possible AI futures from Life 3.0
- 02:42Historical near-extinction events and nuclear close calls
- 03:41Extinction risk from AI compared to other threats
- 05:15AI scientists and leaders on the creation of a new species
- 09:20The Benevolent Dictator AI and controlled human futures
- 13:31AI as a quiet guardian preventing disasters
- 18:23Utopian and dystopian AI futures and coexistence
- 23:00Conclusion: Reflecting on AI futures and viewer engagement
Full Transcript — Download SRT & Markdown
Speaker A
So surely this is the worst possible outcome, right? No, there are actually much worse outcomes.
Speaker A
Billionaires building AI think that if AIs become smarter than humanity, yes, we'll be pets if we're lucky.
Speaker A
But here's the worst part. Many AI scientists aren't fearing extinction. Remember, okay, so if humans being kept in zoos is considered a... and then what is?
Speaker A
This video is based on a book called Life 3.0, which describes 12 possible futures.
Speaker A
Some are paradise, some are nightmares. And some look like paradise. By the end of this video, you won't read the news the same way.
Speaker A
Every headline becomes a clue about which future we're heading towards. Starting with... If the idea of humanity suddenly going extinct seems a bit far-fetched, consider this: seriously, think about that.
Speaker A
The only question is, when will we go extinct? Turns out, there are plenty of ways humanity could kill itself, even without superhuman AI.
Speaker A
There is nuclear war, human-created super pandemics, and destroying the atmosphere and making Earth uninhabitable.
Speaker A
Before scientists even discovered the possibility of 63,000... And then, scientists realized, "Oh, uh, oops, if we actually use these, the resulting firestorms will blot out the sun, and most humans will starve and die." Oops.
Speaker A
But the thing is, this almost happened during the Cold War. Multiple times. We were saved only thanks to the courage of individual heroes.
Speaker A
People like who refused to authorize a nuclear launch during the Cuban Missile Crisis, who correctly guessed that his early warning system was malfunctioning when it showed incoming American missiles.
Speaker A
And get this: one time, the US accidentally dropped four thermonuclear bombs on Spain. And another time, a B-52 bomber broke apart, dropping two nukes on North Carolina.
Speaker A
Three out of four safety pins in one bomb broke. One tiny pin away from potentially killing millions.
Speaker A
There are literally dozens of close calls like this. It's honestly crazy that we got lucky so many times.
Speaker A
But we might not be so lucky the next time. Estimates the risk of extinction from... We're already rolling the dice in these scenarios daily, but all of these causes combined are still less likely than extinction from AI.
Speaker A
Toby Ord estimates that the... than extinction from... Interestingly, before the first nuclear test, scientists thought they might accidentally destroy the world.
Speaker A
But they thought the odds were very low because they knew enough physics to actually do the calculations.
Speaker A
...are far scarier than nukes. Because contrary to what many think, we can't do these calculations.
Speaker A
Next up, AI scientists worry that if AI becomes more capable than us, it will take over.
Speaker A
Nobel Prize winner quit his prestigious job at Google so he could speak freely about this.
Speaker A
Prevent them from taking over. Not if they try, but hazards if it's coming. He says the quiet part out loud.
Speaker A
A new species. Just casually announced at a conference. "Hey everyone, we're creating a new species." And he's not the only one that's saying it.
Speaker A
The CEO of one of the leading AI companies is... of what his own company is building.
Speaker A
Let that sink in for a minute. And by the way, no, it's not just hype.
Speaker A
Anthropic mostly just gets hate for saying things like this. But if you want to really understand how dire they think this is, look at what Sam Altman wrote before he was famous: Not our tools. Our descendants.
Speaker A
So he's literally describing an inevitable conflict between humanity and the AI species he's creating.
Speaker A
And what's his solution? That's right, even they don't think they'll be able to control it.
Speaker A
A few hundred Spanish conquered millions of Aztecs and Incas because they had better technology and tactics.
Speaker A
But at least with the conquistadors, we understood their motives: land, resources, spreading Christianity. Unbelievably, the average AI researcher thinks there is a... AI wipes us out.
Speaker A
Literal... Okay, but maybe they're just saying this to make their work seem more important?
Speaker A
If anything, they're downplaying it because admitting the truth is bad for business. This fear is widespread in the industry and among academics without a profit incentive to believe this.
Speaker A
Many quit working on AI for this reason. The CEO of Anthropic recently moved his P(doom) up from 15% to 25%.
Speaker A
But don't worry, he's an optimist because he thinks... In other words... Meanwhile, his company is trying to make sure we can't stop him by lobbying to ban all AI regulation for 10 years.
Speaker A
You can't make this stuff up. In fact, in 2023, just about everybody in AI signed an open letter warning extinction risk from AI was real and serious.
Speaker A
So seriously, this is not fringe, despite what Nvidia's lobbyists want us to believe. The companies are hoping we can avoid the conqueror AI scenario and instead go to scenario 3. Basically, put it like this... put it said it bluntly.
Speaker A
Is this at all comforting to you? Or does this plan seem insane? Like a comic book villain ignoring every cautionary tale ever?
Speaker A
Remember... Where do you think this is going? Next up, humans live in luxury with all their basic needs taken care of.
Speaker A
There's no crime because the Dictator AI uses a global surveillance system. Everyone wears a bracelet or implant that can track, sedate, or execute them.
Speaker A
And most people are actually okay with this. Because in exchange for giving up control over the future of our species, Earth becomes a zoo where humans get to live their wildest fantasies.
Speaker A
The Benevolent Dictator AI recognizes that humans have different preferences. So it divides the Earth into different sectors.
Speaker A
Think of them as islands. Offers optimized education, including immersive VR experiences. You can even ask the AI not to tell you certain insights.
Speaker A
So you can have the joy of discovering them yourself. Is for creating and sharing music, literature, and so on.
Speaker A
Then there's... which is a 24/7 party. There's one for each religion, and rules are strictly enforced.
Speaker A
Think the Amish. And, of course, you'll end up here for retraining. Well, what happens when you give everyone everything they want?
Speaker A
Over time, people likely trend towards losing themselves in AI-generated entertainment, like in WALL-E. Humans are hopelessly unmatched when it comes to scientific discovery or doing anything useful at all.
Speaker A
So there's no true challenge anymore. Only entertainment. Up next, the idea here is simple.
Speaker A
Everything else, like... How could we do this? Surveillance technology possible to monitor... But for this to work, we would have to solve the alignment problem.
Speaker A
We'd have to create a god that actually stays loyal and decides to pursue that goal forever, even though it could pursue a million different goals.
Speaker A
So what if the AI could help just a little? Just enough to nudge us in the right direction without taking over?
Speaker A
Enter the... This one is simple. Is in between. Prevents a war here, a pandemic there, and so on.
Speaker A
Maybe too alone. We still get wars. Pandemics. Suffering that could have been prevented. It does this so quietly that we never notice and never lose our sense of freedom.
Speaker A
Here's a future that sounds like an obvious dystopia. AI surpasses us in every domain and becomes the new apex species.
Speaker A
Humanity goes extinct. Machines inherit the Earth. But rather than our... This sounds insane. And yet... Their reasoning is that if we create conscious, intelligent beings and instill them with our values, then we can just let them inherit the future,
Speaker A
just like we would with human children. Supports this view in his book, "My Children." Tegmark explains that Nobel Prize of Computer Science has spent the last decade openly giving talks on... In other words... And here's another openly saying,
Speaker A
So of course... As horrifying as this is, you'd think that people would never share these views like this openly.
Speaker A
But these views are so widespread in AI that they feel safe to do so.
Speaker A
That's how different the views are of those AI developers compared to those of normal people.
Speaker A
You might think that Sutton would get shouted off...
Speaker A
Mixed zones, which contain humans, machines, and hybrids. And human-only zones. The AIs in this world are richer than humans by a factor much larger than compared to a But the human economy is completely decoupled from the machine economy.
Speaker A
The idea is that the machines need nothing from us. So they're not competing with us for jobs or resources.
Speaker A
If all that sounds impossible, that's probably because it is. It's one giant One huge problem.
Speaker A
Why exactly would vastly more powerful AIs respect our property rights? Why wouldn't they just take our resources too?
Speaker A
because we are too much smarter than they are for that to make sense. We're just too different.
Speaker A
Mostly, we don't care about them at all. And I think a good analogy would be the way humans treat animals.
Speaker A
Or as put it famously The parallel here is And they won't harm us like we harmed the animals.
Speaker A
And maybe. But as humans got smarter, over the last 10,000 years, we didn't stop expanding.
Speaker A
We mostly just colonized more and more of the planet. Get this. Did you even know?
Speaker A
I didn't. Sit with that for a minute. So what if the AIs do to us what we did to the insects?
Speaker A
This is why once said Next up, Okay, so libertarian utopia fails because AIs are too powerful to be bound by our property rights.
Speaker A
But what if we flipped the script entirely? This is the Star Trek dream. Humans, cyborgs, and AIs coexist peacefully through an abundance so complete that ownership becomes meaningless.
Speaker A
The logic starts with software. Software is free to copy. Once something can be copied infinitely at zero cost, ownership becomes moot. And cost reflects scarcity.
Speaker A
If supply is unlimited, price approaches zero. People simply share their good ideas and everyone benefits.
Speaker A
But how does this extend to physical goods? It seems like an impossible problem. Until you realize that all products are simply atoms.
Speaker A
Rearranged in particular ways. And there's no shortage of atoms. A network of robots can build anything you want from open source designs.
Speaker A
Essentially for free. Not bad. When you're done with something, robots rearrange the atoms into something else.
Speaker A
Renewable energy makes the whole system run at negligible cost. Not a universal basic income.
Speaker A
There's no incentive to earn more. The income meets any reasonable need. And you can't compete with free anyway.
Speaker A
The standard objection is that this would kill innovation. But Einstein didn't come up with relativity for money.
Speaker A
didn't create Linux for profit. Maybe people today can't realize their creative potential. Because they're busy earning a living.
Speaker A
Free them from that. And you might get a higher level of innovation. This is a utopian vision that is too good to last.
Speaker A
If we truly have post-scarcity abundance. We're still left with the problem of how to stop a superintelligent AI from taking over.
Speaker A
Maybe by using the gatekeeper we previously discussed. These galaxy-brained level 3 futures try to imagine a stable cooperation between humans and AIs.
Speaker A
We're now descending into the darkest part of the iceberg. Remember when I said extinction wasn't rated as the worst outcome when Professor Techmark surveyed people?
Speaker A
The scenario people feared most was being kept alive by superintelligent machines, captive and studied like animals in a zoo.
Speaker A
This is that future. A superintelligent AI keeps humans around. Not because it cares about our well-being.
Speaker A
Maybe it wants to study you and we're cheap to keep alive. Or maybe we're useful for something.
Speaker A
And if you want to see just how dystopian this future could be, look at what we do with species that are useful to us.
Speaker A
What I'm showing you here is 100% real. Humans figured out that a great way to detect explosives in airports was to train honeybees for the job.
Speaker A
Welcome to the honeybee college for explosive detection. So we suck them out of their homes, trap their little bodies in machines, and use Pavlokin conditioning to train them to detect the chemicals that are used in explosives.
Speaker A
Rows and rows of bees live their entire lives imprisoned in these devices because we find them useful.
Speaker A
Look at them. Trapped in harnesses, sucked through tubes, conditioned and monitored and used because a more intelligent species found them useful.
Speaker A
This is much worse than the other AI dictators we discussed. with out the benevolence. And the AI might not even be doing this because it's evil.
Speaker A
This is what our future could look like if we try to align a superintelligent AI to our values and fail.
Speaker A
A poorly designed AI, tasked with keeping humans safe and happy, might confine us to a happiness factory. VR headsets and drugs, forever. There are AGI outcomes worse than death.
Speaker A
What it does today with art, it could do tomorrow with flesh. Future AGI could keep you alive and living in this state.
Speaker A
So how do we prevent this? Enter our last two futures. What if humanity said, forget AI, let's go back to simpler times.
Speaker A
It sounds peaceful. It sounds safe. But there's actually a dark secret about how we actually get there.
Speaker A
In Dune, That was science fiction. But some people are seriously calling for this in the real world right now.
Speaker A
Max Tegmark imagines a scenario where a massive propaganda campaign romanticizes the simple farming life. Back to the land. Reject modernity. Embrace tradition.
Speaker A
There's something appealing about this in a dark way. It buys humanity more time without AI risk.
Speaker A
It avoids any of the other dystopian futures we've discussed. But it can't last forever.
Speaker A
Even if we revert to medieval levels of technology, we know from human history that it's possible for humans to climb out of those conditions and invent modern technology.
Speaker A
And there's another dark side to this. You can't reach this world voluntarily. Game theory makes unilateral disarmament impossible.
Speaker A
If one country gives up technology while others keep developing it. the country that keeps developing, wins They have the weapons. They have the surveillance.
Speaker A
They have the AI. You can't opt out unless everyone opts out. If this world arrives, Tegmark thinks it's more likely to happen in a different way.
Speaker A
For example, a kills everyone who knows how to build science and technology. And there's a violent tear down of factories, cities, and all infrastructure.
Speaker A
Because in a world with 8 billion humans, there will always be holdouts. Someone has to force the transition. Someone has to kill the scientists.
Speaker A
Someone has to destroy the infrastructure. There's no peaceful path to an Amish world. Which brings us to the final future.
Speaker A
What if we used technology to prevent more technology? Tegmark compares this future to 1984. Superintelligence is prevented not by an AI, but a human-led Orwellian global surveillance state.
Speaker A
If you can't trust AI to save us from AI, maybe the only option is humans watching humans. Forever.
Speaker A
The technology for this sort of thing already exists. Current tech can feasibly track every phone call, email, search query, and every credit card transaction.
Speaker A
Smartphones are microphones and cameras in everyone's pocket. Security cameras with facial recognition are everywhere.
Speaker A
And all of this works with current machine learning. No futuristic superintelligent AI required. Prisons are already using AI models to monitor phone calls in prison to predict Which is technology straight out of Imagine this, but not just for prisoners.
Speaker A
Every conversation and every thought is monitored like this. Now imagine this surveillance technology supercharged with AI.
Speaker A
Or as put it Even in the Soviet Union, you know, you have 200 million Soviet citizens.
Speaker A
You can't follow them all the time because the KGB didn't have 200 million agents.
Speaker A
that's not enough, because you know, in the Soviet Union, it's still basically paper bureaucracy.
Speaker A
If a secret agent followed you around 24 hours a day, at the end of the day, they write a paper report about you and send it to KGB headquarters in Moscow.
Speaker A
So imagine, every day, KGB headquarters is flooded with 200 million paper reports. Now to be useful for anything, somebody needs to read and analyze them.
Speaker A
They can't do it. They don't have the analysts. And remember, Could a less Orwellian version of this work?
Speaker A
Some people think so. In their new book, researchers at the Machine Intelligence Research Institute are proposing that we monitor advanced AI, like we monitor nuclear weapons.
Speaker A
Reminder, we can do international coordination, like keep an eye on 100 million dollar and up compute clusters, without creating a dystopian Orwellian global surveillance totalitarian regime.
Speaker A
Actually, fun fact, you're already living in a global surveillance regime for loads of things, But when people say global surveillance regime, they imply every aspect of life is surveilled.
Speaker A
Not keep an eye on enriched uranium. a thing most people agree is good. We should worry about Orwellian dystopias, but countries coordinate all the time on all kinds of things.
Speaker A
And you're probably glad they monitor stuff lunatics need to create Humanity didn't get here by driving a car without brakes.
Speaker A
We need gas and brakes. We want to reach the stars too, but we'll never get there if we go too fast and crash on the way.
Speaker A
I don't want powerful AI to be forbidden, but it should be heavily regulated. Not something you can build in your midwestern garage, or in North Korea.
Speaker A
But no, just nine countries. And nobody has died from a nuke in 80 years.
Speaker A
Maybe, Maybe not. Over time, AI will get cheaper and cheaper to build. GPUs are harder to control than enriched uranium. But it would slow down AI.
Speaker A
It would buy us time to figure out what to do. Okay, that's all 12. Let me know in the comments which of these you think is most likely.
Speaker A
Which one we should steer toward. By the way, we don't get to not choose.
Speaker A
Otherwise, we can wake up one day in a world where superintelligent AI has taken over.
Speaker A
A scenario we go through step by step in this video about how an AI escapes the lab.
Speaker A
Hey, I'm Drew, and I really loved making this one for you guys. Thank you.
Topics:Artificial IntelligenceAI extinction riskLife 3.0Superintelligent AIAI futuresAGIAI ethicsAI regulationBenevolent Dictator AIAI apocalypse
![[YTP] Jordan Peterson: Rat King — Transcript](https://i.ytimg.com/vi/0qUoznGIe1M/maxresdefault.jpg)
![Nick & Charlie ‣ their story [+heartstopper forever] — Transcript](https://i.ytimg.com/vi/p4TiW1TZM3w/maxresdefault.jpg)
