How are creative people going to make a living in a world with AI? What does the term data dignity mean? Why is the character of data from Star Trek a useful model for how we could think about what the sources are for any creative expression. Is there a totally different way to think about the economy of the future, and how might this involve mystifying and elevating humans. Welcome
to Inner Cosmos with me David Eagelman. I'm a neuroscientist and an author at Stanford and in these episodes we sail deeply into our three pound universe to understand some of the most surprising aspects of our lives. Today's episode is about the situation we find ourselves in in which AI is so boundless and rich in its creative output. You ask for a picture of anything you want from a good LM like Open AI or mid Journey or Grock,
and it gives you these really superlative results. It recombines elements and concepts in a way that is deeply creative and satisfying. But of course, while it's doing an unexpectedly terrific job of recombining things that are already out there, it didn't come up with the original pieces and parts by itself, there are people, individual humans who generated those drawings and concepts and art forms. It's fundamentally human creativity, not computer creativity, that is the fuel behind it all.
So is there any way to the humans who drove the original innovations or is it simply too late in the sense that maybe everything has already been vacuumed up by the lms and they've digested it all, and now there's no way for human producers of art and ideas to get any meaningful credit for their work. What does this all mean for the creatives in society, those who paint and compose music and write books? What does it
mean for those who produce work now? And for those who are in school now but dream of being human creators in the future. So, for today's podcast, I have the pleasure of speaking with my friend and colleague, Jaron Lanier. Jaren wears many hats. He's a computer scientist, he's an artist, he's a technologist, he's a futurist, he's a critic, and he's a musician who, by the way, owns almost two thousand very unusual musical instruments. Jaren is the godfather of
virtual reality. He started the first VR company in nineteen eighty four, and he's spent his career as a visiting scholar at companies and universities, and since two thousand and nine he's been at Microsoft Research, where he holds the role as prime Unifying Scientists under the Office of the Chief Technology Officer, which gives him the acronym Octopus, which is appropriate given his reach into so many fields. I'll just say that I'm blessed to know a great number
of brilliant people. But even in that crowd, jaren sits near the top of the heap. So today we're going to talk about AI and the future for creatives.
You show me an AI system, I don't care if it's CHET, GIPT, or really anything else. You can think of it in two ways. There's a figure ground and version. That's so. A figure ground in version is when you can look at something and you can swap the way you interpret it almost to an opposite right A famous one is the mcsher drawings, where you might see a field of fish or a field of birds, but each is the negative space of the other. Right now, in this case, for AI, one way you can see AI
is AI is a thing. It's a noun. Whether you think it's alive or not, or conscious or not. Forget that it's just a thing. It's like, the AI did this, the AI did that. Can we regulate the AI, who sold the AI? Who's responsible when the AI did this or that? Blah blah blah. It's a thing, it's a noun. There's another way to think about it where it is not a noun anymore, but instead it's a collaboration of people and there isn't anything there other than the people.
Now when I say this, people are so familiar with treating AI as a thing that they have trouble hearing the version. Sometimes, what do you mean, of course AI is there? Of course it's a thing I just bought. I'm paying for your copilot thing and that's the thing I paid for it, right, and like great, thank you for paying for it. But there is another way to think about it, and it is possible to imagine an AI system as actually the collaborative effort of all the
people who made it. This is particularly true in big data systems like large language models. You can think of them as being closer to the Wikipedia, perhaps than to Commander Data from Star Trek. Although I want to see something interesting about Commander Data. I was just reviewing clips of Commander Data talking and he always introduced himself as
an amalgam of people whose data he was combining. There's a wonderful episode where he plays violin very beautifully and Captain Picard comes up to him so that that was a great performance, and he says, well, really, you should thank the three hundred violinists whose data was amalgamated for me to do this. And what I like about that scene is that he knows specifically the three hundred. They're not a faceless mush of some unknown number, which is
how we do it today. But the writers at the time, and I should fess up, I was in the loop and was talking to people, so I might I might have had a bit.
Of influence on it.
But the idea is that you could know who the people were if he wanted to, and Data could say it seased three hundred violinists instead of well it's some random mush of violinists, you know. So what we've done now is we've mushed everybody together so we don't know who they are, and that's been our approach to information systems, and it benefits those of us who make.
A living from them.
If you're listening to a playlist online from some music service and you get a feed, you don't necessarily know who all the musicians are, but you know the name of the service you're paying the monthly bill to.
So the digital hub.
Becomes more powerful, becomes more known, more prominent, more valuable than the people who are feeding it. So there's this constant economic incentive to emphasize the hub and not all the people who are really the only thing that are there. In the my preferred inversion of understanding the thing, the same thing could be true of AI. We so much want AI to be this emerging entity, even if it's
an evil, horrible one that will destroy us. Because we grew up on the Matrix movies and the Terminator movies and all of these. We want that creature to emerge because that's our childhood. That's almost like our religion. It's the story we grew up with. But if we acknowledge it actually it's all these people instead, then we lose We lose the creature, and that would be traumatic. So when we train the models, we don't keep track of which source documents were involved. So what we really need
to do is eliminate the people. And we have to do that for economics. Can you imagine. I mean, like with economics, we can't trace everything. We don't quite know why a price is what it is. But on the other hand, we could say, hey, buddy, uo taxes, So we have a definite motivation to keep track of the people. With AI, we lose all the people. We just like pretend that the people aren't there. But that's the wrong inversion.
Well, I know that one of the things that you can'tpaign for is digital dividends.
Well we call it.
Initially as an academic research field, it was called data as labor, so you treat your data as if it were labor. And then the name shifted to data dignity. Dignity which was such his idea, such an adella. Not that he agrees with me about everything, believe me, but you know it's that's the term he came up with.
And data dignity is let's.
Say you could you got some result, he said, hey, chachipt, can you write me a Christmas card?
Or whatever?
People do okay, and then they would say, here's your card. By the way, the top twelve sources, I mean there were a multitude and unbounded multitude of sources, but the top twelve were these, and then you could say, hey, could you get rid of that one and do this one instead. It gives you an x ray into what you might think of as the intent of the particular output. And I like that for two reasons. One is, there's
a safety question. So when we have red teams attempt to fool the model to get it to do things that we don't want it to. For instance, can you make a bomb out of what I see here in this kitchen right now? Because I really want to blow this place up. We don't want that to happen, right, But is there some way they can couch that prompt in some kind of a weird thing like asking for a cake recipe instead, but somehow or other, like maybe it indirectly references a movie where there was a bomb
and somehow rather they get it anyway, all right. The thing is, if you can look at the most prominent source documents that were relevant to that result and there's a bomb in there, you've nailed it. So you have an x ray into intent and there's not really any substitute for that, and we lose that when we pretend that there weren't people there, or that they were just
a giant, untraceable mush. And then, of course the other thing is I want I want to have a future where people can be paid for providing exceptionally useful data to AI to incentivize better and better data production FREEI I want that future, and not just the future everybody has to go on universal basic income and be the same and feel useless. So that's the second reason, and that's why it's called dignity.
Got it Now, just to play Devil's advocate for one moment on that, let's say that you ask it to write you a poem, and it says, hey, John Smith was sort of the biggest influence for this poem they got written.
But of course John Smith is just a vessel into which was.
Poured the entire culture, and his output of the poem came from all the other influences on him.
So where do you where do you? Yeah?
So this is a very interesting problem, and this is a fundamental philosophical problem, which is the universe is in a way of giant continuity, and how do you ever draw boundaries? How do you ever say, actually, there's this thing and that thing. It's a very basic, fundamental problem, and it's a trickier one than people might realize unless
they've really confronted it. I'm not going to go into the whole issue of ontology, because it's a big Ye've been working on it for thousands of years and still here what I want to say in this case, I do have an answer, and it's an answer that you might not like and many might not like, which is I think we have to celebrate human beings and elevate them in a way that, if you like, gives them this status of being sources, even though they're always amalgamators too.
And we have to do that on faith, and in a way it's a bit of a mystical idea that there's something about people that's magical and a little apart. You could call it consciousness, you could call it different things. But the reason we have to do that is if we're technologists, we have to define who our beneficiary is for the technology, and if we can't define this special beneficiary, we can't even be coherent technologists. We lose the whole thing.
Become we become random morons. Just jiggling around between ideas is of no meaning. We become memes, we become viral, and nothing's there anymore. So I don't think we have any choice but to somewhat mystify and elevate humans. And so when somebody source document might have internally through other channels, dependent on others, which it always will, I mean, there's of course, it's always true. I still think we have to defer to that person as much as we can.
And if somebody else says, hey, that shouldn't have been their source document. They copied me, Well, we have systems in place for that when it's egregious enough and worthwhile enough, and they can super copyright if they want to. I don't love that, but it's there, and sometimes it needs to be there. You know, it's rough justice, it can never be perfect, but we have to agree that authorship's.
A real thing.
We can't just say that all people are just vessels and there's nothing but this emergent thing and no individual person as a creator. We have to say no people are creators, because otherwise we can't be technologists anymore. The moment you give up people, you might as well go smash your computer because it doesn't make any sense anymore, because it's only possible. Definition isn't serving people.
So let me dig into something about how it's this figure ground reversal of looking at AI as a collaboration of people. So this taps into something that I've been writing lately about this question of whether AI has theory
of mind. And you may know some people have published paper saying, hey, with the theory of mind has emerged, And just as a reminder of the audience, it's you know, theory of mind is being able to put yourself into the shoes of someone else, into the perspective, and understand their beliefs, even if they're different from your own, and understand what they believe is true or not true their perspective so on. Now, this is something humans do seamlessly
and effortlessly all the time. We're very good at understanding, Oh, he doesn't know that piece of information, she does know that, and so on. But the question is do llms do it? So some people have claimed yes. I've studied this point very carefully. I conclude that they do not. Now why do some people write that they do. It's because you give these questions that are probing on. Hey, what if there's, you know, a bag that's labeled with one thing, but
there's something different inside. What does the person believe who's just looking at the label, and what does she believe after she looks inside the bag? And so on and lllls will get the stuff right, and so they say, Wow, they've.
Got theories in mind.
They can understand what it's like to be this person in these different circumstances. The reason I don't think that's true is what I'm calling this the intelligence echo illusion. And what I mean by that is thousands, maybe millions of people have written about this on line. They've written out these deary mind questions, these unexpected things and so on.
So an LM absorbs the statistics of lots and lots of these things, and then when you ask the question, you say, my god, it gave me the right answer. And of course gave the right answer. It's read the damn problem lots of times.
What's interesting, this is the calculation.
I ran recently, is you know, if you look at the common crawl, this corpus of the data that all these large language models crawl and absorb that's, according to my calculation, one thousand times larger than what you could read in a lifetime. And so the fact is, many, many times we'll pose a question to let's say, chat YOUPT, it'll give some answer. I'll think, my god, it's got
theory of mind. It's really done something strange. But in fact, you're just hearing this intelligence echo, by which I mean people already knew this.
You just didn't know that, or you didn't know that other people do that.
Sure, but I mean we're right back in this territory of mysticism and ontology. Because if somebody wished to disagree with you, and that person would not be me.
But if somebody wished to.
Disagree with you, they'd say that all those people were just reflecting things like that anyway, So there's no difference that the AI and the people are trained on are in the same status, the same category. And so what if the AI got to it this way with people in the way, it's still the same thing that people are doing. And why are you making this distinction? Why are you trying to be all mystical in love people?
What's wrong with you? And so I'm just going to declare you should be mystical and elevate people because it's the only thing. Weirdly, this little bit of faith, which I seem irrational, is the only way to save rationality because otherwise, once again we have no beneficiary for technology. You can have science without elevating people, but you can't have technology. Technology has to have a beneficiary. The problem to be solved is serving people, and technology isn't sensible
without a problem to be solved. Science is you can have a science theory without knowing what it's for. You can't have a technology without knowing what it's for. Or maybe I mean one could argue that you can have some kind of underlying, very fundamental technology that might be applied in different ways, but ultimately to actually make any instance of it, it has to be for something. There has to be a human at the end of the
chain for it to be sensible. So I totally agree with you, but you have to recognize that there's a figure ground reversal there, And somebody who disagrees with you might just say, well, the AI and the people are the same and they're both just getting this information. But I still think our only choice to remain rational technologists is to accept that we have to be kind of mystical humanists to frame it, and that's not comfortable for
a lot of people. But I think if you really examine the logic of the situation, you'll come to agree with me.
Oh, I already agree with you, actually, And that's why, that's why I really care about this intelligence echo illusion, which is to say, you're hearing the echoes of people who've said this thing before, but you mistake it for the voice of an AI that has theory of mind.
Yeah, and so why can't you Let's say the AI gets the puzzle with the bag, right, Okay, let's say those now in my preferred future, where you then get a list of the top twelve or twenty five or whatever it is people who contributed. What it might say is I would tell you, but actually there are a million people interchangeably in the first slot here because this thing is so common, and that has to be an
acceptable and honest answer. So like if you say, hey, could you do something like I want a cat that's playing the banjo or whatever, and they'll say, got to tell you a lot of cats out there. It's the Internet, remember, lots of cats, So it really could have been any cat.
In this case, it was this cat. But that's not special.
And so we have to have a fungibility measure, not just an influence measure. And that that's so, this whole future world of acknowledging people is a bit more subtle and complicated than.
I'd like it to be.
But I've never seen anything about it that doesn't make sense or is unachievable.
Right, So one idea that you've mentioned me in the past would be, I can't remember if I got this term digital dividends right, if I made that up some whe along the way. But the idea is that a creator gets paid, He gets a few nickels here and there because on you know, Dolly, it's used his painting as part of the influence.
Yeah.
So the thing about this idea, which currently the term I like is is dated dignity. Many will object and correctly that in a lot of cases, if somebody's output happens to include a picture of your shoe for some image that has a shoe or something, or maybe they used your drone footage of waves for the wave thing or something, what you'll get literally would be very small. It might be quite a bit less than Nichols, right, But that's not the point.
See, the question is whether we think of the.
Future as a linear extension of what we already know
or something really expansive and fantastic and unpredictable. And a lot of the people who tell me that they believe in this radical future where there's a singularity, where this AI transformers reality and a blink of an eye and humans are obsolete, and they're actually thinking in a very linear way because what they believe is the AI as that exists is capable of creating the AI of the future, and that we already know in a sense everything that needs to be known, and all we have to do
is turn the on switch. In future generations won't be more creative than us. Where the final creative, you know? And I think that might not be the future.
I want.
What if the future is one in which there's incredible, creative, productive things happening, and I'm not creative enough perhaps to come up with the examples.
I'll just use the one I used last.
Week, which is, in the future, we have some way of extending our bodies physiologically, so we can fly around an open space in the vacuum and do all sorts of things and propel ourselves. And what if in order to do those body extensions there's some sort of bioprinter we get into. And what if they're aisystems that run that because it's the only way. And what if there's a whole new creative class of people who've contributed data
to that thing. All right, so that's the whole new thing, and I think the best of those people might get rich from contributing data that then is able to be beneficial to people through a large model that their data helped train. And what if there's more and more examples like that. What if there's thousands and millions of tens of millions as humanity expands into an ever more creative, interesting future instead of a future that we already know all about because we're the smartest people.
Who will ever be.
And so in that case, we start to see niches for people who are data creators for AI where they really our novel and they're really making money. And that's what people think so linearly like they like, it's not just TikTok, it's like flying around in space without a space suit, like free your mind. Imagine a future more radical than you assume, yeah.
And maybe that there's a very clear economic incentive there, which is to say, let's say that I have let's say I'm putting pictures of cats online. As you point out that anytime Dolly puts together a picture of a cat, mine is one of millions of pictures. So I can't make any money that way. As a creator, I am therefore incentivized to do things that are really new, to really push the matteries of what had been done in terms of art, poetry, whatever it is.
Let's say a little bit more about that incentive, because that's very important. There's a question of whether the network dynamics dominate what actually happens over the network or not. So when network dynamics dominate, what you see is phenomena of virality and memes, because those are network effects where it doesn't really matter what the meme is. It doesn't necessarily matter what goes viral. Sometimes silly things do, sometimes interesting things do, but it can go either way, all right.
In a real economy, when somebody loses money on some stupid NFT or meme, stock or something, that's the network effect dominating. Usually typically the reason our world hasn't collapsed totally. Is that when people spend money, it's not on total nonsense, but it's at least slightly on something that's real of some use. Okay, And in that world of a real economy, there are incentives to make things better and there's a
chance for competition to be meaningful. But when the structure itself dominates rather than the thing the structure is channeling, then you end up in this make believe world. And on digital networks, the make believe world is always about being the lucky one who gets in on the virality. Your startup turns into the hub for this giant thing that benefits from everybody else's work, but you benefit more from them. You get to have TikTok or Instagram or something.
In this world, you have the meme stock, you have the meme video, you have whatever, but you can't have a predict who will get that. There's a randomness to it, which means it's not fundamentally productive. In the future, we need to suppress the network's own influence on what happens on the network and have it back off and let the content itself, whatever that might be, dominate, all right, and then that creates reality that creates incentives for improving reality.
So the most fundamental idea is that you have this data structure that gradually changes and settles into something that can do some work. It happens automatically or semi automatically, or not explicitly programming it. Now, there are a lot of algorithms that have that kind of settling effect, but the grandaddy of them is called gradient, and there's a wonderful mathematical dispute about who should be credited with it, but it's either Cashi or Ramon, So it goes way back.
And it's sometimes thought about as being like walking on a landscape. So when we walk up the landscape, we just think we're walking at lungitude and latitude and it's some elevation.
But in AI we're dealing with very very.
High dimensional spaces, which is a concept some people aren't familiar with. But any rate, let's just imagine we're walking on a landscape, and so you want to descend and you want to find a comfortable place, like where's the place you can go? That's really low if you don't know an advance, you don't have a map or an overview,
so you start wandering around. Now the dangerous if you just say, oh, here's a nice depression, I'll go here, it might not be as low as some other one that would be available, which would be a better solution, And so you have to take some precautions and start to have some kind of way of saying, well, I'm going to kind of go around a little bit so I don't get caught in the place that actually isn't
ideal or something like that. You have to have a strategy, okay, And so the point is not to get caught at some halfway solution.
You want to really kind of jump around.
Now with AI systems, the whole point like if you want to, let's say I want you to have a diffusion model that puts a cat on a motorcycle and with whatever. Okay, so if the cat's riding a motorcycle, the easy solution is just to make a cat or a motorcycle, getting both kind of hard. And so the easy solution that isn't useful is very similar to getting stuck in one of these intermediate dimples that doesn't.
Go as low as what you might really want.
You wanted to send to a better one where you're getting both the cat and the motorcycle is you have to do some extra little thing to broaden your scope. So what I was going to say is this sort of this sort of strategy is fundamental. It's the very core basics for what we do these days with AI. Their many, very many versions, but this way of doing things is the fundamental trick of AI. But when we
put those AI algorithms out in society, we forget. We just say, oh, well, if the AI algorithm can make money by getting kids addicted to vain attention seeking and low attention spans, best, fine. But it's not. That's one of these dimples. It's not good for society, it's not good for the kid. And so the nerdy way I would put some of this is to say the same discipline we need to get the algorithms to work at all needs to be applied to the way we deploy
these algorithms out in the larger society. And that might be a nerdy way of saying something similar to what many of us have been saying about trying to make something that serves people better.
So wait, just so I put a cap on that isn't the case, then that you agree that it would incentivize creators to be more creative if they were doing things that hadn't already been done by a million other people, because they get no money for that. But if they do something that is really new, then they can. There's a potential to make a lot of money from that.
So one of the fundamental claims of people who like market economies is that they foster creativity. The type of competition that happens in a market economy is said to insteadivize creativity. I am on the side of the people who believe that market's foster creativity. I just I can't prove it, but I can almost. I think I can almost prove it. I mean, I mean, look around. We're in Silicon Valley, for God's sakes, give me a break. Of course, of course they do. Of course markets do.
But the Internet. When the Internet is allowed to be dominated by network effects, by virality, and by memes, then it no longer does, because then you're chasing the benefits intrinsic to the Internet, which are not in themselves creative, but just this kind of random, lottery winning kind of a thing.
Right.
That's why I think, Yeah, providing an economic counterbalance where the less MEMI you are, and the more original you are, the more that has economic value. Way to reverse the network property, because I'm not sure how it's going on its own right.
So one of the ways I summarize data dignity is to say, whenever we have an opportunity to create a new creative class instead of a new dependent class in the future, we should do that. So if there's a bunch of people who might be put out of work because robots are driving the cars or whatever, we should try to find some other group of people who might be creating in some way that they could be paid for, where now we don't expect them to be paid like
we should be creating creative classes. There's another caveat or qualification to this idea, which is that we can't expect everybody to be creative, and we can't allow creativity to become the measure of human worth or value. But I do think whenever you can create a new creative class, it's imperative to do so.
That was my conversation with Jaron Lanier, computer scientist, inventor, thinker, writer, innovator, critic, technologist. The issue we talked about today is that people who are creative, the painters, the writers, the composers. They have been justifiably worried ever since AI really hit its stride just a couple of years ago, because the stuff AI can create is stunning. But maybe with the right economic structures,
AI can launch us farther into human creativity. This just requires incentivizing things so that a creative doesn't do just another thing that's been done before, like another cat video, but instead something genuinely bleeding edge. In a data dignity economy, you will benefit only when you reach out into the uncharted wilderness to do something novel, and when the AI draws from that and other people benefit from that, then
you can make a living that way. This is a way to reward an exploration of the frontiers that speeds up the whole human race because it encourages people to be creatively productive, not reproductive. So as we think about the world, we've suddenly shot into a world with AI disrupting every industry. I hope today's conversation will remind us of a fundamental truth. AI isn't a standalone creation. It's a mirror. It's a mosaic of countless human contributions. Every
line of code that it writes. Every line of poetry, every beautiful piece of art is there, all reflections of human creativity and human effort. So the choice to view AI as a collaborative endeavor rather than an autonomous entity, it's not just a philosophical stance. It's a call to honor the humans behind the machine and to ensure that they are valued in a future shaped by their contributions. Go to Eagleman dot com slash podcast for more information
and to find further reading. Send me an email at podcasts at eagleman dot com with questions or discussion, and check out and subscribe to Inner Cosmos on YouTube for videos of each episode and to leave comments Until next time. I'm David Eagleman, and this is Inner Cosmos.