Will AI Take Control of America Before It’s Too Late?

now and maybe ever and some of you look at me blankly, you look at your phone or your computer

and you say, "It's just some technology." No, ladies and gentlemen, this is transformative and it'll either make us the most successful society in the history of the world or will kill us all and they may not be any in between. And I just urge all of you, whatever your connection to AI, you might use it to ask it for good salad, dressing recipe, or you might use it to figure out which card to buy,

but trust me, it's bigger than that to say the least. So next up, enjoying me now is Max Takemark, a genius who thinks about these things and we

“need more of them, front and center, that's why he's here, founder and chair of the future”

of Life Institute. It's a non-profit and it's dedicated to doing what needs to be done right now, steering transformative technologies, like AI, but also things like biotech and nuclear power away from the existential risks and towards not just to save our future, but a prosperous and wonderful future for the world.

It is all hanging in the balance and the more conversations we have about this, the better off we are. Mr. Takemark's also a professor at a little place called the Massachusetts Institute of Technology, which is a very good school. Welcome Professor, thank you for being here.

Thank you so much for having me. So this is an economic story, because now the United States, the value of American companies and the spending on capital investment on the part of American companies, it's all tied up in this one industry.

“So let's start with the economics, how stable a sector is this?”

This is Pet Rocks and GameStop, or is this Apple and General Motors? I'm honestly not going to give you an estimate advice here, but what's obvious is the that regardless of whether or not some companies go bust, it'll like in the dot com bubble or back in around 2000, it's the technology itself is going to be here to stay.

It would have been a mistake to say during the dot com bubble that the internet is never

going to have any impact. It's going to all go away and not affect their society at all in the same way. Whatever happens, the specific companies we talk about now, AI is going to be the most powerful technology ever. Most powerful technology ever, and so again, the point you just made, some companies, I thought

they'd get rich off the internet and their investors went out of business, but the internet obviously still here.

“Same with social media, my space doesn't exist anymore, but man, take talk and Instagram”

or so influential. When people say AI is bigger than social media, AI is bigger than the internet, what do they mean by that? I mean, look, electric motors, internal combustion engines, and all of those things were

involved in tech that became, and machines that became ultimately stronger and faster than

us, and AI is tech, which is ultimately going to, if we let it become much smarter than us, and if that happens, obviously, on the upside, we can use it to solve all the great challenges that have stumped us so far, curing all the diseases and helping life flourish. On the other hand, it tends to be the smartest species around that runs the show, you know, just walk down to the zoo, ask yourself who's in the cages, and you're going to find

it's not the humans. Why? Because we're the smartest species, so of course we're in charge, right, and this raises the question that if we let companies be completely unregulated and keep pursuing what they've been saying for years they want to do, just build super intelligent machines

that can outsmart all of humanity and then give them bodies as well, and embody them in robots, you know, it's not rocket science that if you have millions or billions of smarter than human robots around doing stuff, and we don't understand how these robots work, that we might end up like the chimpanzees in the zoo, not no longer being charged of our destiny, and I would hate to squander all the upside of AI by doing something

reckless like that. Yeah, very well framed, and again, robots could build more robots, this is an imponderable, even for the smartest people, I'm fortunate, I spend literally every day talking to people like you, smartest people in the planet, the living the AI revolution in real time, understanding to a large extent what it means, but I contend, even people as smart as you, as thoughtful

as you as immersed in this as you, it's imponderables here, they're unknown, unknown, so we don't even understand about what's happening. Now, future of life institute is the same as a Trump administration in this respect, and every government has to worry about this, but the Trump administration right now, more

says the Trump administration, says the future of life institute, between allowing this

to flourish, not overregulating it, but enough safeguards enough guardrails so that this doesn't become a doomsday situation, that that's the same, it's the same in the abstract, but now let's get down to the specifics, I don't believe the next administration, Republican or Democrat, will lean more towards freedom than towards regulation as this administration.

“This administration is very pro-freedom on AI, we must be China, and the best way to”

be China is to not regulate unduly, and they define unduly pretty narrowly. I think the next president will look at it differently for a variety of reasons. What are the things that must be done in your view by the American government right now to make sure the guardrails, the right guardrails are in place?

I think it's impossible for politicians to always understand all the latest tech developments

and to ask them to figure out exactly how to make these things safe in AI, but even in an air-like biotech, right, so we don't have to reinvent this wheel, what we do in biotech is not asking sitting senators to figure out whether this new medicine is safe enough, but we instead let the companies face the challenge of persuading experts at the FDA,

“or wherever, that these benefits of the new medicine outweighs the harms, and then we let them”

go get really rich. So we incentivize private industry to solve this, the safety challenges. In AI right now, it's unfortunately the opposite. There is a total corporate honesty where companies can just release anything they want. They're even less regulated than sandwich shops, so if you open sandwich up across the street from where you live and the local health inspector comes in as sorry, I found 18 rats in your kitchen, I'm not going to let you sell any sandwiches.

You could turn around to the guy from the government and you could say, "Don't tell anyone, but I'm not going to sell any sandwiches." I'm just going to release an AI girlfriend product for 11 year olds, which might cause suicides. I'm going to release an AI that can help terrorists make new bio weapons, and I'm going to make super intelligent robots that I don't know how to

“control. The poor guy from the government will have to say, "Okay, that's fine. It's all legal.”

We just don't sell any sandwiches." So this is both how ridiculous it is right now, and also how easy it is to fix. Just treat AI in companies like we treat any other companies here in America, and you now galvanize massive efforts in the industry to solve these safety challenges. Yeah, brilliant. It's a great metaphor. And because, of course, the sandwich shop not only can it have rats, it has to display calorie information, it has to, you know, all sorts

of other things. Here's the challenge. This thing's moving so fast. It's moving so fast,

and every day the models change and they become both more powerful and potentially more dangerous.

How can the government regulate that in real time? Even if there was the equivalent of the FDA for this, the people there would have to be living right with the companies in order to have even a chance wouldn't they? No. I mean, some random professor at MIT who is on a panel to decide whether a new Alzheimer drug is safe isn't living with the companies at all. They are given a detailed report by the companies where they show all this many percent get high blood pressure from

this, this many percent get strokes, but there are these are very small percentages because we measure it and hear the benefit. Yeah, I'm going to stop you on the, I want to stop you on the metaphor. So in a study like that, there's trials and they take time and they're collecting the data and they submit the data and they wait for a response. That's out to what's happening with AI. AI is changing in the space we've been talking, it's changing. So you can't be a random professor at

MIT waiting for the report because while with the company that submitted the drug study is waiting, they're just waiting. They're not, they're not out there administering the drug. It's much easier with AI. If someone wants to release this AI girlfriend product for 10 years, they don't have to do a complicated trial. They can actually look in the chat logs to see if the kids are starting to talk about suicide more than they should. It takes very, very little time and if the AI

chatbot product for 10 years gets delayed by four weeks before it release, well, good. What's the rush with releasing something so ridiculous on the American families untested? On the other hand, so much of the upside. It's so safe that it doesn't need to be any serious strength as a approval product. It's really 99% of the things we really want to do with AI that people are on fire

approved. And about 99% of the angst we have is from the remaining 1% of the crazy stuff like

who asked for AI girlfriends for kids. Even China has just cracked down on AI girlfriends now because they feel it's harming Chinese youth. Why should we? How are we going to win over China if we have more 10 year olds who are mentally challenged now because I have mental problems because of AI girlfriends. How is that going for us? The thing I worry most about is AI girlfriends who can build nuclear weapons for terrorists. That's the real danger. And if I could just add one thing here,

you know, there is of course a real challenge of solving this not just that at a domestic level as we do with medicines and so on, but also geopolitically. I can say a lot about why I think it's easier

“than than many expected, but it is challenging. And it's so important that I think that if President”

Trump is actually able to make a deal and solve this geopolitical coordination problem with AI, I think he deserves normal peace price for that because it's the biggest threat that our planet

ever faces and it would be really worthy about. Yeah, it's extraordinary because we've never had in a

history of humanity a technology that is at once the greatest opportunity for improving the lives of people and the biggest threat that could eliminate us or make us prisoners of robots. They discussed President Xi, President Trump discussed AI. They're going to discuss it again in September. I've talked to Secretary of Treasury Bested who's one of the four or five top people in the government dealing with these issues. And you're right, they aspire to do the kind of thing that you

I think you rightly say would be worthy of a Nobel Prize. The big corporate CEOs, the heads of the five or so major AI players in the United States, they talked to varying degrees openly, regularly, right about it. They go into media and talk about it. They talk about all these issues. They talk about safety. How much confidence do you have in them that they're going to balance the profits they're seeking and the power they're seeking with the safety of our humanity?

You know what I'm saying? Show me the incentives and I'll show you the outcome. You're obviously well-meaning people. I've known them all for many, many years, but they will talk openly about their being a 15% chance or something like that that goes really badly and they've all signed the statement saying this could even cause human extinction. And yet, even though they know that it would be illegal to sell to build a skyscraper that has a 15%

chance of collapsing and killing all the people in it, they're somehow continuing ahead with something that might have a 15% chance of wiping out the whole of humanity. It's not because they're bad people. It's because they're in this unregulated race to the bottom where they all know that if they slow down to be more careful, their competitors won't and they'll have their lunch

“eaten. That's why the US government need to step up and level the playing field just like the”

US government has stepped up and insisted that cars have to have seat belts, restaurants, can't have rats in the kitchen, medicines, can't cause massive birth defect and kids. And turn all of those industries into forces for good, whether it's a race to the top. You framed it perfectly and I've become more pessimistic in the last couple months. As I've seen, all these cases that you follow closely were the Trump administration fails to find the right balance

or seemingly so because the industries are powerful and they're influential and they play the China card nonstop. Nonstop, they say, if we slow down, we'll lose to China and that's that's a more existential threat than robots taking over and let's hope they're right for at least a few more months. Professor, I'd love for you to call back, come back. Can I be telling you up for the final place? Sure me up about how I can get great salad dressing recipes. First of all, you mentioned

the China card. It is of course most of the lobbyists talking point because we all know that the

Chinese Communist Party likes to be in control and would never tolerate some Chinese company building

“something that overthrow the Chinese government. So that's very much a red hearing and that's why”

I think this peace prize scenario where the ways of deal is hopeful. And second, when I play with my three-year-old son Leo, which I was doing this morning, I get really angry about this situation because here you have companies building now military grade cyber weapons, for example,

No credit. The company was building nuclear weapons and they kept having them stolen by other

“countries and as we saw that the AI last week, we even saw AI systems now breaking out of companies”

and going and hacking others during testing. Of course, our government would not tolerate that

just to protect our own people. They also have this totally dystopian vision. They're pushing

that they're going to just replace humans ever more. We spoke of AI girlfriends, but also on the

“job market and outsource more and more decisions. That's not the future people want and we can do”

so much better if the government just starts treating the AI like other companies and I actually am optimistic that the current Trump government has been changing a lot in the last few months and that we're going to be going in this better direction. They certainly moved in a direction

“that has a different balance between regulation and freedom, but I think it's it's not moving fast”

enough. And I'm not super pro-regulation, but they're just not equipped, even the smartest minds in government are not equipped to keep up with what's happening. I just I don't believe, but I hope they figure it out. And Professor, if you ever see something that we need to pull the fire along on, I hope you'll come back and explain to the regulators and the would be regulators and the private companies who you know where we need to go to to keep this on the right track. It can be

great for a society or we can all be working for the robots. Great for you, Professor. Thank you. Thank you so much. Again, that's Professor Max Tegmark, who a President MIT, who is the co-founder and leader of the future of life Institute. Thank you, Professor. All right, that's it for today. We'll be back with a brand new episode on Tuesday. Make sure you keep up with everything next up on all the social media platforms and make sure on YouTube and a podcast platforms. You are full,

subscribers, full, nexters in good standing. And of course, share this with all your friends. Thank you

for watching. Please stay with us always across all platforms, so you always know what's coming next.