Menu
Sign In Search Podcasts Charts People & Topics Add Podcast API Blog Pricing
Podcast Image

80,000 Hours Podcast

Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

27 Jan 2026

Transcription

Chapter 1: What historical factors have influenced the need for democracy?

0.031 - 15.884 David Duvenaud

The reason that states have been treating us so well in the West, at least for the last, let's say, like two or 300 years is because they've needed us. And in particular, because allowing freedom and like private property and basically self-determination has been the most effective recipe for growth. Life can only get so bad when you're needed.

0

15.904 - 23.456 David Duvenaud

That's the real key thing that has been keeping governments aligned. And that's the key thing that's going to change. A lot of citizens would end up just being sort of like full-time activists.

0

23.476 - 40.694 David Duvenaud

And they might feel like they're forced to because if their only source of income is something like UBI, then the entire game going forward for economic advancement is do some sort of activism to convince the government to give your group more UBI. And those same resources could be used to simulate maybe like millions of much more sympathetic, morally superior virtual beings.

0

41.174 - 46.66 David Duvenaud

And so it'll start to be seen as this like irresponsible use of resources to keep like some sort of like legacy human around.

0

50.605 - 54.292 Rob, Luisa, and the 80000 Hours team

Today, I'm speaking with David Duvenaux, professor of computer science at the University of Toronto.

54.933 - 72.968 Rob, Luisa, and the 80000 Hours team

David is a co-author on a somewhat recent paper called Gradual Disempowerment, which makes the slightly counterintuitive claim that even if we manage to solve the AI alignment problem and have AIs that faithfully follow the instructions and goals of the group that's operating them, humanity could nonetheless end up losing control over its future and end up with a pretty bad outcome.

72.948 - 89.27 Rob, Luisa, and the 80000 Hours team

The paper got a lot of reactions, it's fair to say, with some people saying that it really put its finger on an underrated issue and others thinking that the scenarios painted were really unlikely and other people arguing that they were likely but not even necessarily undesirable. I'm kind of a bit unsure where I come down myself. So thanks so much for coming on the show to discuss it, David.

89.45 - 90.171 David Duvenaud

Oh, it's my pleasure, Rob.

90.932 - 100.305 Rob, Luisa, and the 80000 Hours team

So let's imagine that we have managed to make big breakthroughs in AI alignment, you know, maybe around 2028. How is it that nevertheless things could end up trending in a negative direction?

Chapter 2: How might AI change the relationship between citizens and governments?

122.134 - 128.841 Rob, Luisa, and the 80000 Hours team

So we basically can trust the AIs to follow the goals that we give them. How does humanity begin to become disempowered?

0

129.479 - 146.256 David Duvenaud

So the first one is economic and people basically losing their jobs and becoming unemployable. And I think a lot of people get off the bus here and they say this is the classic lump of labor fallacy. And they think that we think that there's only like some finite fixed set of jobs. And if they're automated, then that's just the game over for humanity.

0

146.236 - 163.134 David Duvenaud

And of course, they're right to point out that on the margin, there's always more valuable work to be done. And you can always employ somebody to do valuable work at some wage, basically. And comparative advantage will mean that there will always be some sort of profitable trades on both sides that mean that in principle, humans will be employable at some wage.

0

163.555 - 179.054 David Duvenaud

But the problem is that this breakdown is for two reasons. One is just transaction costs can mean that it's not worth the hassle for the machine companies or whatever is running the economy to employ humans. Like humans are pretty unreliable and kind of a pain to employ for a lot of reasons.

0

179.114 - 192.192 David Duvenaud

And we don't want to hire, for instance, like 12 year olds, even if they're going to work for a dollar an hour. In fact, that's illegal. And that's one of the sort of structural forces that we also expect to be operating is that humans are just going to be this unreliable sort of scary thing to involve in anything important.

192.372 - 198.501 David Duvenaud

And it'll be it'll be seen as like irresponsible to involve them in important decisions once machines are like better alternatives.

199.038 - 215.964 Rob, Luisa, and the 80000 Hours team

Okay, yeah. So I agree that economists have many good reasons to think that mass unemployment from artificial intelligence is going to potentially take quite a while, that machines might have to be significantly above our level before humans just won't be able to get jobs at all. But I think... It's more of a delaying game.

216.444 - 232.204 Rob, Luisa, and the 80000 Hours team

At some point, we're going to have machines that are far faster, far more reliable, and potentially can do all the things that we could do for less than it would even cost to feed a human and keep them alive. I mean, of course, by that stage, businesses will have reoriented their entire – all of the factories will be redesigned to be built around AIs.

232.624 - 244.46 Rob, Luisa, and the 80000 Hours team

Office work will be being done by AIs at such a speed that it's barely even possible for a human being, if they're involved, to keep up with what's going on. And then it won't even be that humans are not able to help. It's that involving them would actually be a negative. It would slow things down.

Chapter 3: What are the implications of economic disempowerment due to AI?

667.268 - 677.809 Rob, Luisa, and the 80000 Hours team

Yeah, I guess it's already the case that I feel, I mean, I think it's like getting close to the point where I spend 50% of my time or something in the office, like basically speaking or interacting with AIs of one form or another.

0

678.09 - 686.627 Rob, Luisa, and the 80000 Hours team

So it's definitely true that kind of their beliefs effectively or kind of memes that are propagating through LLMs are kind of propagating through me and then affecting my actions.

0

686.708 - 697.747 David Duvenaud

Yeah, and I think people are really recognizing that the beliefs of AIs or the constitutions of AIs are a key front in cultural battlegrounds. It's just like people used to fight over Wikipedia to try to set the narrative.

0

697.787 - 718.959 David Duvenaud

And I think now if you want to set the narrative on some controversial topic, if you can really control how ChatGPT frames it, that's going to set what the default cultural answer is. In a sense, this is business as usual. People already notice that economics or cultural forces or like geopolitical forces end up pushing us towards outcomes that I think no one would endorse.

0

718.979 - 731.233 David Duvenaud

And we argue that the development and proliferation of smarter than human AI is going to make those forces even stronger and remove some of the safeguards that tend to keep our civilization serving human interests in the long run.

732.134 - 740.003 Rob, Luisa, and the 80000 Hours team

What are some of the ways that like current economic or cultural forces are pushing us towards less than desirable outcomes today?

740.388 - 758.818 David Duvenaud

Maybe a simple example is just clickbait short-form video content or something like that, where maybe the consumers realize it's not the best use of their time. They kind of regret spending a lot of time on it. And the producers also know what they're doing, and they're making clickbait, and they know it. But if they want to make their educational long-form content

758.798 - 774.124 David Duvenaud

content like such as this, then they get punished by not having as many views. And now each of these content creators having their own aligned AI doesn't solve this like global problem of the market just incentivizing for this sort of not very helpful to humans cultural content.

774.34 - 790.982 Rob, Luisa, and the 80000 Hours team

Yeah, I think the cases that stick out most in my mind are we still have violence and still have war, despite the fact that that's negative some and not really in the endorsed interest of any particular group relative to a negotiated agreement that would get you to the same outcome without having to go through the violence first.

Chapter 4: How could political disempowerment occur in an AI-driven future?

1904.78 - 1920.777 Rob, Luisa, and the 80000 Hours team

for human beings to live and have a good time and grow their food and so on is going up because technology is advancing. We're figuring out more, like how to squeeze more and more, you know, AI, more and more productivity, more whatever we value out of kind of each square kilometer of surface on the earth.

0

1921.357 - 1932.829 Rob, Luisa, and the 80000 Hours team

And so it's becoming potentially more expensive to keep humans alive than it was before, at least in terms of what we're giving up. And then possibly some humans won't actually be able to afford that increasing price because their income won't be going up as fast as that.

0

1932.809 - 1949.991 David Duvenaud

Yeah, exactly. So the scenario I have in mind is that I have my maybe like one acre of land or my like big luxury apartment with me and my family. And, you know, we've made our peace with kind of opting out of the economy, but we have our like little sort of commune or whatever that we're happy to live in, in like unimaginable luxury and wealth in some senses.

0

1949.971 - 1969.178 David Duvenaud

And the government or the rest of the economy or something starts to view this as sort of like criminally decadent, that this small group of like humans, like maybe 10 or 100, are using this entire acre of land and like this amount of energy and sunshine to keep like this sort of these small brains working for no particular benefit but their own.

0

1969.158 - 1988.382 David Duvenaud

when the same resources could be used to simulate maybe like millions of much more sympathetic sort of morally superior on whatever axis like virtual beings. And so it'll start to be seen as this like selfish, as you say, like high opportunity costs, sort of irresponsible use of resources to keep like some sort of like legacy human around.

1988.682 - 1995.691 Rob, Luisa, and the 80000 Hours team

I'm just guessing that like by this point, surely there has been some kind of agreement that humans are going to have like some fraction of the earth.

1995.771 - 2008.007 Rob, Luisa, and the 80000 Hours team

We're going to be sort of grandfathered in, like either we've been killed or there's going to be some sort of agreement that we're going to be allowed to have some section of the earth in perpetuity in order to support ourselves while the robots go off. And like the earth is very small in the scheme of the entire universe.

2007.987 - 2027.195 Rob, Luisa, and the 80000 Hours team

And it's a lot easier for AIs and robots to go and use all of the resources in space. While the opportunity cost in an absolute sense of setting aside, you know, half of the Earth's surface for human beings to do their thing on would be very big. In terms of proportion of all the available resources in the universe or even just the solar system, it's basically completely negligible.

2028.196 - 2034.265 Rob, Luisa, and the 80000 Hours team

So why do the AIs care so much to kind of squeeze that last bit of, you know, space and energy away from the human beings?

Chapter 5: What are the potential consequences of gradual disempowerment on liberalism?

4866.072 - 4878.252 David Duvenaud

Totally. And I want to say, I think this is a tragedy. Like I'm a huge liberalism enjoyer. And I mean, maybe you could say, well, yes, things are working out for you. But I guess I just feel like in general, it's very easy to, like liberalism is also just this very fragile thing.

0

4878.272 - 4892.755 David Duvenaud

And it's sort of this like amazing accomplishment that the West and just managed to create these norms of like, let the other guy choose his religion, let the other guy get rich. It's like a very sort of fragile thing that like, I think even today, we should still try to protect if we can.

0

Chapter 6: How does universal basic income (UBI) influence societal competition?

4893.562 - 4903.153 Rob, Luisa, and the 80000 Hours team

Okay, so what are the ways in which you think liberalism might be less competitive as a system and less attractive, less appealing a way of organizing society post-AGI than it is today?

0

4903.873 - 4923.535 David Duvenaud

Yeah, so I think it'll be a less desirable way to organize society for a few reasons, but the main one is just the zero-sumness of UBI. So, you know, right now when we all create our own wealth, it doesn't really hurt me if someone else creates their own wealth, like sort of directly from resources. But in the world where we're all just like living in some like apartments advocating for UBI,

0

4923.768 - 4943.177 David Duvenaud

To the extent that the UVI pie is fixed, then we're really just like a bunch of baby birds like cheeping and whoever gets food is like less food for the other guy. And then this also erodes the sort of like pluralism of values because the government's going to have to have some way of deciding like who gets resources. And so if they end up having any opinions about like what

0

4943.157 - 4948.026 David Duvenaud

way of life is more valuable or needs to be subsidized more or whatever, that could be a threat to you.

0

Chapter 7: What are the implications of AI constitutions for governance?

4948.066 - 4961.35 David Duvenaud

So you kind of have to argue like that guy's way of life is less deserving of resources than my way of life. And so now the government is forced to decide like whose sort of values are most like Well, the fact of who gets subsidized. That's the main effect.

0

4961.85 - 4983.373 Rob, Luisa, and the 80000 Hours team

I see. So we almost have to imagine a hypothetical society in which no one can make anything. There's no economic production occurring, at least among this group. There's just a fixed endowment of resources that they happen to have found and run into a certain amount of food, a certain amount of houses, and all of that. And they've got to figure out how to organize themselves. I guess...

0

4983.353 - 4994.675 Rob, Luisa, and the 80000 Hours team

It's not necessarily desirable for me for you to have free speech and to be able to advocate for yourself all that well, or to be able to like educate yourself and become more and more powerful and influential because it is completely zero sum.

0

4995.357 - 5004.535 Rob, Luisa, and the 80000 Hours team

The more influential you become, the more you'll be able to advocate for getting stuff that is literally like food out of my mouth or like money out of my bank account. Is that the main thing that has changed?

0

5004.515 - 5005.777 David Duvenaud

Right. So that's the big thing that's changed.

Chapter 8: How can we effectively address the risks of AI development?

5005.817 - 5024.885 David Duvenaud

And there is, of course, a way in which this might not be zero-sum, right? Like, if humanity manages to convince the AIs or whatever, the government, to give a larger UBI overall, then that is, like, the sort of normal positive-sum thing. So that might not be, like, a slam-dunk argument. The other thing, though, I think has been that we haven't had to fear domination by other groups very much.

0

5025.025 - 5035.322 David Duvenaud

So we've had strong property rights, right? Like, I'm not afraid that, like, Elon Musk is going to, like, literally take my stuff, even though he could, like, raise a private army or whatever, right? We have very little variation in reproductive rates.

0

5035.964 - 5052.334 David Duvenaud

It's okay that the Amish live nearby because even if they're having more kids than whatever other population, that's not going to matter over the course of 50 or maybe even 100 years. And then maybe another thing is just like the rough egalitarian in terms of sort of intelligence and power level of people.

0

5052.435 - 5070.685 David Duvenaud

So, I mean, there's definitely like very meaningful variation amongst humans in terms of just like raw smarts. But like people often say, like von Neumann somehow didn't take over the earth. Right. And he might have like wanted to. And these are all reasons why it was sort of fine to just let other people become more powerful in the past that might change in a big way.

0

5071.525 - 5084.97 Rob, Luisa, and the 80000 Hours team

So to push back on this, what other sort of positive sum dynamics might continue to exist? I mean, I guess in as much as people all think that the thing that I want to do is the morally right thing, then they might – and there are other people who like also are pursuing that goal.

5084.99 - 5098.096 Rob, Luisa, and the 80000 Hours team

Even if they disagree about specifically what that is, they might still be in favor of pluralism because that might help – they might think that that will – lead them to like converge on a good answer. It's more like in as much as people just have like raw preferences, they're like, I just want to benefit me and you just want to benefit you.

5098.497 - 5107.776 Rob, Luisa, and the 80000 Hours team

And there's like no real way that we can reconcile it. There's no way we can come to agree on what the good thing to do with resources is, where it's like you have less interest in allowing the other person to advocate for themselves.

5107.891 - 5127.47 David Duvenaud

And I just want to say, I want to reiterate that we should be very, very scared of accidentally crushing liberalism and this positive sum sort of world that we've created. If I somehow, through this podcast, accidentally contributed to mistakenly just making people think, oh, you're in a zero-sum game, you need to fight more. That would be a huge tragedy.

5127.59 - 5140.903 David Duvenaud

I really want to err on the side of caution here and say, probably until things change a lot, we really think liberalism is this precious, awesome, amazing accomplishment that we need to foster. So you were saying like, what are the positive sum dynamics that might still operate?

Comments

There are no comments yet.

Please log in to write the first comment.