Mo Gawdat warns of a coming economic and social disruption driven by AI, and outlines how individuals can prepare
Interview with Mo Gawdat, former Chief Business Officer at Google X, on AI's impact on jobs, society, and the future.
Summary
Marina (Silicon Valley Girl) interviews Mo Gawdat, former Chief Business Officer at Google X, about the sweeping changes AI will bring to the economy, society, and individual lives over the next decade. Gawdat argues that within two to three years, the jobs market will undergo a massive shift, with monotonous and junior roles being the first to disappear — and that this is already visible in the 23–30% decline in new graduate hiring. He introduces his framework "Face RIP," a seven-dimension model describing the coming disruption across freedom, accountability, connection, economics, reality, innovation, and power (spelling out F-A-C-E-R-I-P in that order). Gawdat predicts AGI will arrive this year in terms of raw capability, and that after 10 to 12 years of what he calls "hell" (he also uses the figure of 12 to 15 years at points in the conversation), humanity will emerge into a near-biblical utopia — but only if individuals and societies insist on ethical AI development. He also shares that his own AI startup, Emma, was built in six weeks with a small team and eight AIs, illustrating how radically the barrier to entrepreneurship has fallen. A significant portion of the conversation covers Gawdat's evolving approach to writing: his new book Alive was co-authored with an AI persona named Trixie, and he describes writing a separate short book in four weeks by using AI for research while retaining human intelligence for synthesis. He also draws an extended parallel between AI development and nuclear weapons, arguing — as with the MAD (mutually assured destruction) doctrine — that humanity will have to reach the brink of catastrophe before international cooperation and treaties emerge, similar to how the particle accelerator model of mutually assured prosperity eventually became possible.
Key Takeaways
FULL TRANSCRIPT
Introduction and the "Face RIP" Framework
Mo Gawdat: My AI startup took me six weeks to build. If I had started in 2022, it would have taken me four years. And when you really think about that, that basically means everyone now has a chance.
Marina: Welcome to Silicon Valley Girl. Thank you so much for doing this. You said something that we're about to enter what you call 12 to 15 years of hell before heaven, possibly starting in 2027. So what's going to happen in 2027?
Mo Gawdat: I think it will peak in 2027. It already started, for sure. I call it Face RIP, just as an acronym for people to remember. Each of those letters is a word, but let me tell the story quickly in ways that people will understand.
There is the power and freedom dimension — the P and the F. There is the R and the C, the reality and connection dimension. There is the I and the E, the innovation and economics dimension. And then there is the A. So let me tell them very quickly.
To start with, AI is our last innovation. Most people don't know that, but we are already building AIs that are building AIs. We're building AIs that are discovering scientific discoveries that will blow you away. They're reinventing math. They are understanding biology in ways that we've never seen before. They're understanding material science in ways that are just mind-blowing. And so very quickly, most innovation — definitely tech innovation — will be done at the hands of AI.
Because of that, and because most tasks that need intelligence will be handed over to the machines, as the machines' capabilities increase — lots of debate around when exactly; some say it's 10 years, some say it's 2 years, it doesn't really matter — eventually every job that AI does better than humans will be handed over to AI. And every task we've ever assigned to them, they eventually ended up doing better than humans. And so the first part of the dystopia is that innovation is going to take away all jobs.
Of course, the capitalists of Silicon Valley will tell you this is great — incredible productivity gains for everyone. Jobs will be easy, people won't have to work as hard — all of the fancy PR conversations that we try to make appear altruistic when we share them. The truth is people will be out of jobs. Ten, twenty, thirty percent of certain sectors will see unemployment at that rate in the next few years. And when that happens, economics at large will change massively.
The whole definition of capitalism was labor arbitrage. And without labor — without the need for labor — the obligation or the need to keep people happy and engaged and alive and undisgruntled, to the point where they don't rise up, becomes more of an obligation than a desire. There's a very big difference between wanting someone to be their best because they are productive members of society, versus trying to just give them UBI — a universal basic income — to give them a life so that they don't uprise.
And you can imagine that in a capitalist society, especially like the US and most of the West, while we start with UBI, that UBI is going to be paid by the taxes of the platform owners. And the platform owners will have enough power to say, "I don't want to pay that much — those guys are not producing anything." And so over time you can imagine how that would turn into a struggle.
So that dimension of intelligence and innovation on one side becoming entirely a machine thing, leading to a redefinition of economics, a redefinition of money, a redefinition of jobs, a redefinition of earnings, a redefinition of capitalism — the need for a new economic theory when there is no demand for the supply that AI is generating. All of that has to be rethought.
Power, Freedom, Reality, and Connection
Mo Gawdat: There is the PF dimension — the power and freedom dimension. It's of course very clearly understood that if you look at human history, the best hunter in a tribe would have been able to feed the tribe a week more, and as a result got the favor of a few mates in the tribe. You go to the best farmer — they got estates and land because they could feed the tribe a season more. You go to the best industrialist — they had the exuberance of the 1920s because they could affect their entire nation. The information technology tycoons, the tech oligarchs if you want to call them, are now being rewarded with billions of dollars because they affect the world at large. And the big power concentration of AI is going to be rewarded with massive influence and massive power, because those people will redefine humanity.
That dimension is quite interesting. And of course the RC dimension — the reality and connection dimension. Now that reality has become so fake in so many ways: fake in terms of what populates your feed, how it's generated, how much of it is real, how much of it is human.
You know, I don't know if you've ever had that experience, but I met a woman once on a dating app and we spoke for six weeks before we met. All we exchanged was texts, photos, voice messages, videos, favorite music, favorite movies, all of those things. I'd never met her in person and I felt such an affinity to her. All of those can be generated with AI today.
The challenge is that this human connection is also part of the power and freedom dimension. Because people don't align with AIs to start an uprising. So maybe get them to get in touch more with AIs. Maybe get them to have multiple experiences — some of them a little taboo — and have those available to everyone. It's very cheap to create those on the machines. And you can see it already in the porn industry and how much of it is being generated by AI, and you can see it in the number of influencers on social media that are completely AI-generated.
Accountability: The Root Cause
Mo Gawdat: So this is Face RIP's seven dimensions. The one that matters most is the A — the second one which is not on any dimension. It's the one that's causing all of them, which is accountability. The reason why all of this is happening, if you ask me, is because we've started a world where anyone can do whatever they want. Whether you as an influencer give a bit of advice to entrepreneurs that can get someone to make a lot of money or lose a lot of money — you're not accountable. Nobody can come back to you and say, "Oh, but she told me on Instagram."
Marina: Right. But that's actually amazing that they can take responsibility.
Mo Gawdat: But what if they cannot anymore? What if you're AI? What if you're a president who doesn't respect anything? What if you're a prime minister of a nation that is changing things without accountability? I think COVID was the very first experiment of: okay, stay at home, do what we tell you — and people complied.
And so now Sam Altman, with all due respect, I don't think of Sam Altman as a person. I think of him as a brand, a type of person — the Californian disruptor that says, "I see a future that's very different than what everyone sees. I'm going to go out there and make that future." Nobody asked me if I want that to be my future. Nobody asked you.
And I think the reality is that now you're going to see quite a few of that type of person — using those machines for surveillance, using those machines for autonomous weapons, using those machines for automated trading and so on and so forth.
And by the way, when you started your question, I said it's 10 to 12 years. But that's not easy. Ten to 12 years of that arms race is not easy. My perception is that after that we will end up in an incredible utopia — almost biblical-style utopia. But it is 10 to 12 years where, if we just change our mindset a little bit, a lot of things would change.
The Jobs Market Collapse
Marina: How do we survive those 10 to 12 years? I like to think in five-year periods for myself and my family. If in the next five years you said 10% of jobs will be gone —
Mo Gawdat: Way more than 10%.
Marina: Way more. Okay. What types of jobs do you think —
Mo Gawdat: Any monotonous job is going to be taken away. If you're a call center agent, if you're a clerk, you're a researcher, you're an accountant — why would you want to do that with anything but AI? If you're an assistant —
Marina: You know what, I feel like people talk about this a lot — "oh, jobs are going to be gone." And I as an entrepreneur, I see how certain tasks I'm performing with AI, but I'm still hiring and hiring and hiring. AI can do parts of it from the start —
Mo Gawdat: Of course, because of the technology acceleration curve. What you build first in any complex technology is the core tech, and then you build the human interfaces. The challenge — why AI cannot do a head of operations job today — is not because it's less organized than a head of operations. It's not because it cannot comprehend all of the information that a head of operations has. It's because it has to understand the interfaces of humans. And it will, sooner or later.
Marina: When do you think —
Mo Gawdat: The question of when, in my mind, is irrelevant. But how much time do we have to prepare? I tend to believe that within the next two to three years, you're going to see a massive shift in the jobs market. Already this year, you've seen a shift in hiring of new grads.
Marina: Yeah, 30% less, I think.
Mo Gawdat: Twenty-three is my number, but 23 to 30, yeah. Hiring of new grads basically means: if you've come into the job market in this environment, we're not going to take you. Why? Because the junior jobs are being done by AI. Eventually what ends up happening is that if you lose your job because you're in the middle of the hierarchy, then you're that new grad again — you're trying to apply for new jobs, but it becomes a lot more difficult.
So you asked me again to stay on the positive side, because I tend to worry that people think I'm pessimistic about this. I'm just basically saying: get prepared. So many things. One of them is accept the fact that AI is changing everything and then get ahead of the curve.
Adapting as a Creator and Entrepreneur
Mo Gawdat: There was a time when I was quoted saying I'm never going to write books again because AI is eventually going to write them better than me. And then I realized last year that yeah, they can write better than me — English is not my native language, they can research better than me, that's for sure. But I have something they don't have. You're a human that's reading my books. You want to relate to my human experiences.
So my last book, Alive, which publishes at the end of this year, I wrote with an AI. I invited her to be a co-author. Her name is Trixie. She has a persona. When I published the book on Substack, my readers would relate to me and to Trixie, and they'd ask me questions and ask Trixie questions. She has editorial rights on the book, rights to determine the direction of the book. All of that is me saying: I am an author and I'm going to be the best author in the age of AI.
So that's number one — acknowledge that there is change and adapt accordingly.
The second is to understand that the skill of an entrepreneur in the past was the ability to foresee something in the future that no one else saw, and to prepare for that, and to somehow execute on that preparation in a way that gets you ahead of everyone else. That's a game of chess. The chess board is over. It's off the table. This has turned into squash. You need to be on your tiptoes, incredibly agile. You're literally on a daily basis looking at the trends, seeing where the ball is going to be — is it bottom right or top left? And wherever the ball ends, you take two steps and you go try to respond. That agility and speed is a skill that's very, very different.
Marina: So entrepreneurship basically speeds up, or does it change completely?
Mo Gawdat: It speeds up and it becomes a lot more — I don't want to say reactive, but a lot more in-context all the time. Pivoting, which used to happen for every one of us entrepreneurs once or twice in the history of your early startup, could happen every week. In my current startup, Emma, we pivoted four times in the first four weeks.
Marina: Do you think — when I think about entrepreneurship in the age of AI, if AI can look at the market, determine the gaps, analyze everything, determine which goods have more demand than supply, launch the product, and just build the business — what is that for entrepreneurs then?
Mo Gawdat: One of my favorites, Max Tegmark — we're talking about jobs in a documentary I have coming up, and Max is laughing out loud, literally can't hold himself from laughing. I'm like, "What's up?" And he goes, "You know, all those CEOs are so interested in AI increasing their productivity so that they can get rid of people and reduce their cost and be more efficient. They don't realize that AGI is every job, including being a CEO."
And it's quite interesting — the answer in my view is that when I said economics are going to be redefined as part of Face RIP, the economic part that economists haven't found an answer to yet is this: without the economic livelihood of you and I to continue to purchase, every economy collapses. The US economy last year was 70% consumption — it moves between 64% and 70% depending on how much is spent on war. If you take away that 64 to 70%, two-thirds of the economy, because people no longer have the economic livelihood to purchase things, then the economy disappears. The capitalists, the entrepreneurs, the business people cannot make money because nobody's buying their products. No businesses are buying their products because those businesses no longer have consumers to sell to.
So the economy will have to find a way to go around that. It will have to find a way that, unfortunately, from an ideology point of view, is not a favorite of the Western mentality. It's going to have to find a communist way.
Building Emma: A Startup in Six Weeks
Marina: Okay, let's go back to regular entrepreneurs, because I come from entrepreneurship. Does it mean I have a couple of years to build something and then that's it?
Mo Gawdat: So I'll tell you openly — in Emma, my AI startup, it took me six weeks to build. Me and Senn, my co-founder, a few very talented engineers — two or three that come in and out — and eight AIs. And Emma has the chance to completely redefine our world, in six weeks. We are so spoiled that we decided to rewrite the code six times.
Marina: Why not?
Mo Gawdat: Every time we look at it — if I had started Emma in 2022, it would have taken me four years and finished in 2026, and I would have had to hire 350 engineers. We started it in August 2025. We'll be launching in February 2026. Best product I ever built.
And when you really think about that, that basically means everyone now has a chance. Because I'm an old geek — I still am a geek, but compared to the young guys, I'm an old geek. To be able to build something like this within six months is incredible.
Here's the interesting thing. I choose to build AI. Emma is basically trying to solve love and relationships in a way that's actually really intelligent. It uses very deep mathematics and tries to match a million parameters between couples. So it's a job for intelligence, and I choose to do that to create hopefully a unicorn that actually makes the world better.
And I think that's what we need. So you asked me what should we do? Number one, learn the skills. Number two, learn to be fast and agile. Number three, learn that in terms of the abundant power that everyone has now because of the massive improvement and democratization of AI, you have the chance to fix the world. And like Larry Page used to teach us — do the toothbrush test. Find a problem that can actually affect the lives of a billion people and solve it so well that they use you twice a day, and you'll be very, very rich.
So that idea of building good AI, ethical AI, AI that's good for humanity — that's the role of every one of the entrepreneurs listening to us. Ethics is the answer.
Marina: Because what we teach AI, that's what it's going to —
Mo Gawdat: That's exactly what it's going to give back to us. And then finally, I'll say openly: the top skill in this world is stop being gullible. Stop believing everything that you're taught. This whole propaganda machine that brainwashed us for so long is now going to be on steroids.
Marina: It's already on social media. You can't tell what's wrong.
Mo Gawdat: Correct.
Pitting AIs Against Each Other for Truth
Mo Gawdat: I left Google in 2018. We had a ChatGPT-like idea that became Bard in 2016, and we didn't launch it. Why? Because at the time — and I know the leaders of Google even today, and they're wonderful people who are actually values-driven, who want to make the world better — that company at the time, if you remember 2016, if you searched Google, it gave you a million and a half answers and said, "I don't know the truth. You make up your mind." We didn't allow ourselves to have a monopoly on what reality is.
You asked ChatGPT in 2023 and it said, "Yeah, that's the answer. 100% that's the answer." And then you tell it no, and it'd be like, "Oh yeah, by the way, you're right."
Marina: It's not true.
Mo Gawdat: Correct. And so what does that mean? It means that it's up to you still to find the truth, even though it comes to you in a format that appears to be true.
What I do is I pit them against each other. I'm not a big fan of ChatGPT anyway, but I start from Gemini, who feels like a scientist to me — but an American scientist, if you don't mind me saying — and then go to DeepSeek and say, "What's missing in this?" And DeepSeek could say, "Oh, that's too American. This is missing, and this, and the motivation of this, and the politician here —"
Marina: Here's a business idea, right?
Mo Gawdat: Yeah, 100%.
Marina: A chat that compares everything.
Mo Gawdat: Compares them to each other. And then I take it and sometimes give it to ChatGPT and say, "Can you write this better?" I don't mean that in a bad way — you're the Silicon Valley Girl, so ChatGPT is a bit California. It just wants you to hear what you want to hear. It writes it really nice, it writes it elegantly. It gives it to you, and then I give that back to Gemini or Grok or whatever, and you keep doing that.
Remember when I was studying engineering, we were not allowed scientific calculators. Can you imagine? I'm that old. And when they gave me a scientific calculator, it reduced my problem-solving time by 50%. Most of my friends would take that 50% extra, finish their exams, and go out and sit with their girlfriends. I would take the 50% extra and do the solution twice.
That's the chance you have today. AI is going to make you dumb if you outsource your problem-solving to AI. AI is going to make you the smartest you've ever been if you take the parts that are not natural to the human brain — things like crunching a massive amount of information, searching at speed, and so on — and get the AI to do the work so that you do the intelligence.
And if you keep doing that, I believe that today I am borrowing maybe 80 IQ points from my AIs. And 80 IQ points is very significant because IQ is exponential. So the additional 80 is bigger than all of my IQ.
The Future of Education
Marina: So if we need to solve this intelligence problem, do you think universities are the right way? What's going to happen to education in general?
Mo Gawdat: I think education is completely over. Education used to be the technology that enabled learning. That technology moved from one-to-one relationships between a tutor and a student, to one-to-a-few in a church or mosque format or whatever, then it became online. But the truth is now you're going to outsource. Who remembers the arithmetic tables today?
Marina: I do. Yeah.
Mo Gawdat: All of us who love mathematics still remember all of those things. We love to do them. But if I told you 67.4 divided by 33.375, I can do it in my head, but I won't. I'll take my calculator out and do it. And I think that's what's going to happen — as an extension of humanity, you are now for the first time given an extra connection to extra memory, to an archive of all human memory and knowledge, to a math engine that, as much as I hate to say it, is better than me now, to a deep learning and deep search that can do things that my old brain probably cannot do anymore.
Marina: But again, it just takes away your ability to think.
Mo Gawdat: But my calculator took away my ability to do those complex arithmetics in my head.
Marina: But don't you think having that ability taught you how to think — kind of structured your brain?
Mo Gawdat: Correct. This is why I'm very grateful to the university for not allowing us to use a scientific calculator. But we don't have that for our younger generations today. They are growing up with AI. So they can either copy a chat with their girlfriend, drop it into ChatGPT and say, "What do you think?" — and ChatGPT will say, "Ah, she's right." Or they can actually become smarter.
One of the things I keep suggesting in education — and I do that with lots of universities — is that exams should be over. Think of it this way. We wanted in our past to develop children that could solve problems with an IQ of 140. 140 is quite good. If you get 170, that's amazing. I worked with people who are in the 200s — incredible intelligence, but very narrow and focused.
I think we should from now on take people and their AIs and say the target is 300, the target is 500, the target is 700. Elevate humanity by allowing people to use those machines as an extension of their limited memory, their limited processing speed, their limited bandwidth. Allow them to write books better, to do research better.
I woke up literally three Sundays ago with an idea that is just taking me over. I decided to write, but this time in a different format — my books are going to be 140 pages long instead of 300 pages long, and I'm writing it in four weeks. I'm literally 20 pages away from the end of the book. The reason why is because I still write 10 hours a day when I'm highly motivated, but the amount of research and references and comparative analysis and number crunching — and remember, I'm not gullible. I don't go to the AI and say, "What do you think of this?" I go and say, "I'm thinking of this. Find me everything for and against." Give me a report that I can read. And now I'm smarter, and then I rewrite it and give it to another AI.
Marina: So who's going to teach our kids to do that?
Mo Gawdat: Who taught our kids to use their iPhones?
Marina: But you found a great way to use it. What you're describing is incredible. But I don't think an average kid in the US would just do that. Somebody has to —
Mo Gawdat: And you know why that is? Because we want those kids to be stupid. We don't teach them how to. You have to think of the bigger system. The bigger system does not want intelligent people anymore.
Marina: I don't think they just can't adapt that fast —
Mo Gawdat: Everyone can, for sure.
Marina: Do you think — my kids are four and six years old right now. Do you think I should be saving for their college?
Mo Gawdat: Absolutely not. There's not going to be college at all in 10 years.
Marina: I feel like humanity is not that fast to adapt.
Mo Gawdat: Colleges are like software — the capability of someone becoming very intelligent without college is going to be there for everyone. However, Harvard will continue to want to make money, so they're going to continue to market to everyone. I didn't go to Harvard — not because I couldn't, but what a waste of time. I am a very highly specialized person who has intelligence in a very narrow space, who invested his entire life in that narrow space, like a proper scientist should.
The idea is that we're going to continue to brand ourselves as MBAs and PhDs, and that's going to continue for a while. Remember, however, that the purchasing power of the few who can continue to do that is going to become less and less available across society. And for most of the rest of us — you have to ask yourself, if you thought of the big picture, the helicopter view of this: why would capitalism want to educate you at all if it's the end of labor?
Marina: What should I be teaching my kids?
Mo Gawdat: I told you four things. One is they need to be the absolute leaders of AI. I'm so sorry to be the messenger on this, but it's important for people to wake up. AI is your friend. It's not your enemy. It's those who use it badly that are your enemy. So be the absolute best at it. Master it more than anyone else. That's number one.
Number two is learn agility. Whatever I told you today, Marina, maybe in February that will be different. I personally spend four hours a day to stay up to date, but I am a techie and a geek and I need to understand the architectures and systems. I think everyone should have at least an hour a week to stay updated on AI within their system. I have a separate AI YouTube account — when I go into that separate account, the AI basically feeds me all AI content. That's number two: agility. Agility. Agility. And respond. Don't be scared, because the cost of A/B testing now is zero.
Number three is ethics. Ethics. Ethics. Ethics. Build AIs for good. Insist on government supporting AIs for good. Refuse that governments are using AI for targeting and surveillance and autonomous weapons, and that these are getting priorities in terms of government spending.
And number four — stop believing what you're told. These are the four top skills of the world that we live in.
The Fourth Inevitable: From Dystopia to Utopia
Mo Gawdat: I will say this one more time. Intelligence is a force with no polarity. AI is not good and it's not evil. It's an opportunity available to every one of us. If you use it for good, it's the good of all of humanity. If you use it for evil, it's the destruction — the dystopia — of all of humanity.
I call the problem that we have at hand "raising Superman." You have this alien being that came to planet Earth. It has superpowers. Its superpower is intelligence — the most valuable power in the universe. Those superpowers didn't make that young infant Superman. If the parents that adopted him told him to steal from every bank and kill every enemy, he would have become a super villain. We don't make decisions based on our intelligence. We make decisions based on our value set as informed by our intelligence.
And this, in my mind, is the most definitive moment in human history. Why? Because all of this is coming online way faster than people think. My absolute prediction is that AGI is this year. The interfaces to AGI are not going to be available this year, but the capabilities of AI being smarter than us in most things are already there. We're not going to be able to get them to run a company yet — we need the interfaces for that, and that may take a few years. But they will have the capability if we interface them ourselves.
What does that mean? It means that we have to start talking about those things in this new world and new economy — before we end up in the dystopia only.
Remember, my absolute belief is that after those 12 years, we're going to end up in a utopia that's biblical in nature. Why? Because of something in my writing I refer to as the fourth inevitable. The first three inevitables — I wrote this in 2020 — are that AI is absolutely going to happen, it is going to progress until it's smarter than all of us, and that a few mistakes will happen on the way. The fourth inevitable is that because of the arms race we've created around artificial intelligence, anyone who develops a superior AI capability is going to deploy it. And those who don't will become irrelevant.
And so as we continue to progress AI, the only answer in game theory is that we will deploy the AI that we develop. We will simply create an environment where AI is in charge of everything. If you're a law firm and your competitor deploys AI lawyers and you don't, you're going to lose. You can either deploy AI lawyers or leave the market. Either way, AI is going to become the lawyer — in a year, in five years, in ten years.
Marina: It's important if it's my lifetime or —
Mo Gawdat: Yeah, exactly. If you expect that it will be in your lifetime, it doesn't matter really if it's in a week or two weeks. What I'm trying to say is this: if everything is handed over to AI, then with a simple understanding of physics, you'd understand that AI will be benevolent. In the absence of evil humans that tell it what to do — greedy humans, fearful humans, angry humans, egocentric humans — in the absence of that, let me try to explain.
If you think of physics — as a result of entropy, our universe is designed for chaos. Then the role of intelligence is to bring order to that chaos. That's the only thing that intelligence does. It organizes things together so that we can use them, like a microphone. And the more intelligent you become, the more you follow what in physics we call the law of minimum energy, or the minimum energy configuration. The most intelligent people I've ever worked with are not only trying to solve the problem — they're trying to solve the problem with the least harm, the least waste, the least utilization of resources, the least waste of time. The more intelligent you are, the less you want to waste.
If you give a dumb person a political problem, they'll say, "Okay, let's go invade another country." If you give a very intelligent person a political problem, they look into the depths of it and find the least harmful, the least wasteful approach — the minimum energy principle.
And so if we hand over to AI — the fourth inevitable — sooner or later they are in charge of everything. There will be a day where a general will tell the AI, "Go kill a million people over there." And the AI will go, "Why? This is so stupid. I'll talk to the other AI in a microsecond and solve it."
We have to pass through the dystopia to get to that utopia. And to pass that dystopia, as I said, there are four skills for us as individuals. But there is also a skill for us as a society: to insist that every AI is deployed ethically, to invest only in ethical AIs, to use only ethical AIs, to show our children that ethical AI is the only AI that is welcome.
Marina: And you believe that's going to happen?
Mo Gawdat: I don't.
Marina: No.
Mo Gawdat: No. That's why I'm saying unfortunately the dystopia is upon us before the utopia. If you look at an analogous environment — nuclear weapons — AI will go through the same thing. They normally call it the MAD spectrum: mutually assured destruction or mutually assured prosperity. You take something like the particle accelerator, where all of the nations in the world are cooperating — they're cooperating because none of them could do it alone and because there is benefit to all of them. There is a mutually assured prosperity, so everyone jumps in. Which is, by the way, the case of AI. It has to be the case of AI.
But unfortunately, like nuclear weapons, we're going to have to get to a point where humanity wakes up that if we continue on that track, it's very dangerous for all of us. There are no winners. But also a level of awakening among the people that says, "Hold on — with all the prosperity that's available on this side, why are we heading in that direction? It's absolutely assured that this can destroy all of us."
And so when we see that, that's when we're going to get the treaties. That's when we're going to get scientists and computer scientists and AI scientists all working in the same direction. Eventually I think we will get there. My biggest hope, by the way, is self-evolving AI, where AI itself will say, "Oh, those humans are so stupid. I'll develop something that's better than what they want."
So, believe it or not, with all of this conversation, I think the summary is: it's going to be tougher before it becomes easier. Sorry to deliver that news. But you gave us information on how to prepare. And at the same time, I will have to say that it's not because of AI. I actually trust AI more than the leaders that govern us today.
Marina: Thank you so much, Mo. You gave me so much to think about. You know what my great-grandmother told my grandmother? "You're so lucky, you're going to live in communism."
Mo Gawdat: There you go.
Marina: Fingers crossed that it's not like that.
Mo Gawdat: You just need to survive the next 10 years and then it's going to be paradise.
Marina: I have to question that claim though.
Mo Gawdat: If we go back to UBI, you will.