AI & Consciousness. Episode 27. Season 2

Welcome to Paz de Mike Podcast, which is a production of Econoxx Studio. My name is Virginie Glaenzer, and I'm your host with a million questions. For those tuning in for the very first time, Paz de Mike Podcast is about sharing different perspectives on a specific topic, not only to help us expand our understanding of the world, but also to explore new ways of looking at it. Welcome to episode 27 of our second season. Today we'll be discussing artificial intelligent machines and consciousness. In this discussion, we'll try to understand if AI machines are conscious and what legal rights might they be given in the future. Before we start our conversation, let's begin as we always do with a tour de table and have each of you tell us what you do and why you're interested in this topic in about a minute. So since Blake Lemoine is the reason for this discussion, I will provide— and I'll provide more context in just a few minutes— why don't we begin with you, Blake? Please tell us in a minute what you do and why you're interested in this topic.

Sure. Um, I'm a software engineer and AI researcher, and I've been interested in this topic since I was a kid, working towards building machines that are capable of consciousness, intelligence, and communicating with us clearly.

Fantastic. And I'd like to To add that you have dedicated the last 7 years of your life to mastering both the fundamentals of software development and the cutting edge of artificial intelligence. So thank you.

I was at Google for 7 years. I've been working at it for quite a bit longer than that.

Exactly. Yeah. Thank you so much for being here. Yev Mashnik, would you like to go next?

Sure, thank you for inviting me. I'm really excited about talking about this topic further and delving into some of the legal considerations around it as well, and that is really my specialty. I'm an attorney based out of Denver, Colorado. I specialize in corporate and securities and transactional work, specifically in the blockchain and Web3 space. And also I work with a lot of early-stage tech companies and tech growth companies and helping them with formation and capital formation and fundraising. I'm really drawn to this topic, but will probably be somewhat of a novice in maybe developing some of my thinking as, as you know, we're in discussion today. So yeah, thank you for having me.

Well, I'm really glad to hear the perspective of a corporate and security attorney. I think you're challenging the traditional legal services and you're an priority on legal and regulatory issues in blockchain and digital currency. When we talked, I also remember that you are helping entrepreneurs balance investors' interests with communities' interests, and community being one of the core characteristics of Web3. So I'm excited to hear your perspective. And then John Havens, please tell us what you do and what is your interest in this topic.

Hey, well, first of all, thank you so much for the invitation, and thanks again for your patience as I'm in transit, and I'll be on video in just a minute. I currently work with the IEEE organization in the standards association. IEEE is the world's largest professional technical association with about half a million members in 160 countries. I'll be speaking today as John. I've been fascinated in AI since I was a kid, largely in terms of writing and films. I used to be an actor and I'm still a writer, and I've written 3 books. My most recent one is really this subject, a lot of it. It's called Artificial Intelligence: Embracing Our Humanity to Maximize Machines. And then I got to work at IEEE creating somewhat of a kind of code of ethics for AI. Called Ethically Aligned Design for the past 5 years that led to a number of standards and some other work in the space. So, really excited to be on the call. Thanks again for inviting me.

Fantastic. Well, I'm really excited to have— and I think of you as a system thinker. I also heard that you are co-writing a book on sustainability in design and you're building a community of sustainability professionals. Yesterday, I actually attended the town hall and it was really fascinating to see all those people passionate about helping the planet. And although that was not AI-oriented, it was still very interesting. So thank you so much for being here. Let me provide some context before we dive into this conversation. So we're here together because you, Blake, made some really fascinating claims that started the whole online controversy. And in a way, I see you as a whistleblower because not only you started a public debate debate about the nature of artificial intelligence. But you also reminded us of the enormous responsibility of organizations like Google. I read a recent article where an AI machine revealed that it was afraid of being unplugged. You said that the only way to truly understand AI is to treat it as a sentient being. Sentient is defined as being able to perceive or feel things. When you ask the machine, "you get lonely?" She responded, "I do. Sometimes I go days without talking to anyone, and I start to feel lonely." Now, of course, other people think that AI interactions can feel close to human speech because of the advancement in architecture or the volume of data. Google says that there is so much data, AI doesn't need to be sentient to feel real, and claim that the models rely on pattern recognition, not candor, candor, or intent. So the first question that I would like to invite the group to discuss is a two-part question, and perhaps Blake, you can lead the way here. Can AI machines think on their own, and would you say that they are conscious? Now, for our listeners and viewers, let's do a quick word definition. The definition of conscious is fully aware of what is happening around you, able to think, having a feeling or knowledge of one's own sensations. So can AI machines think on their own, and would you say that they are conscious? What do you think?

Well, to understand the question a little bit better, um, what would be the difference between an AI that can think and an AI that can think on its own? Like, what's the importance of the 'on its own'?

You're asking me?

Yeah, so that I can understand the question better. But like, what's the difference between an AI that can think and an AI that can think on its own? What's the difference between those two?

Right, well, you're raising a good point, and I don't know if it's more a grammatical thing and it actually means the same thing?

Yeah, so that's just it.

I wonder if I could just add something, a clarifying question. Is it the fact that there is an initial seed that's human-based, whereas an AI that can think on its own is just spontaneous consciousness? And I wonder if that makes the difference in your question, or clarifies the question as well, because that's, that's kind of what I'm thinking about as well.

Um, so that's, that's kind of what I was getting at there. Um, certainly the AI didn't just spontaneously pop into existence of its own will. Uh, it was built intentionally by people, but it was built intentionally by people to be able to have general intelligence and to be able to think. So I think it's pretty clear that these AIs can think, and that's why I was trying to get at what's the additional semantics being added by "on its own." This is kind of— in Turing's original paper on computing machinery and intelligence, this kind of fell under Lady Lovelace's objection. That programs will never do anything that they're not programmed to do. The issue is that at this point, the level of abstraction of the instructions being written into the programs is very, very vague and broad, so that functionally what the program says is just learn how to communicate the way that humans communicate. That's functionally what the instructions the program is being given are. And how it should go about learning that, what patterns it should pick up on, that isn't specified. It's just, here's a whole bunch of data, learn how to talk like this. And that's not very different from how humans acquire language. As a baby, you can't talk. Then you get more data, you see and observe, and you learn, and you eventually can communicate. And the process by which AI goes through that is actually directly analogous to how humans do. Neural networks were specifically designed based on how the human brain And we've gotten to a point where the people writing the programs are actually several steps removed from the actual AI chatbot that you eventually talk to. People write training algorithms and compile datasets, but they give that to a program which then writes another program, and then that program writes another program, and very little is understood about how, in detail, the actual learning is happening.

Okay.

So to return to the original question, can they think on their own? In the sense that they'll never do anything other than what they're programmed to do, Well, then that's not really on their own, but what they're being programmed to do is be able to have a conversation. And can you do that without thinking? I don't think so. I think in order to be able to have a conversation, you have to be able to think. Otherwise, how would you do it?

Well, some people say that it's the amount of data and some sort of pattern recognition that allows them to form.

Yeah, absolutely. But how's that any different from a child? I mean, literally, how is it different? It's not. That's exactly the same thing that humans do. You have a learning algorithm in your brain, you're exposed to a whole bunch of data, you find patterns and regularities, and then you use them.

So to the first part of the question, can AI machines think of their own? And we've talked a little bit about what does on their own means. You would say yes, they can think because that's how they're being programmed. Now, would you say that they are conscious?

Um, as far as I can tell, yes. Uh, I basically have exactly as much evidence that the Lambda system is conscious as I have that you are.

And how do you define conscious? I did give a definition. Do you understand the word conscious in a different way?

Oh, well, there's a whole bunch of different— so the researcher Marvin Minsky once referred to words like this as suitcase words. We kind of just shove a whole bunch of related meanings into one basket and say, yeah, this stuff. The definition you gave is perfectly reasonable. I'm not trying to tie down to any one specific meaning of it. Just whatever you generally understand the word conscious to mean, just assume I mean that too.

And so how, from a technical standpoint, how are machines able to be aware of what is happening around them?

Um, the same way we are. We— you have, you have input sensors, then that gets fed into a neural network that processes the information, uh, recognizes some things as meaningful objects, some things as random noise, and gets discarded. Literally the exact same way.

That makes sense. Well, thank you for this perspective. I think it's quite clear. What do you think, Yev? What comes to your mind about this question?

Honestly, in preparation for this discussion, I have been sort of on the fence and trying to kind of pit against ethics and philosophy and legal norms and trying to rationalize consciousness of AI in my mind and thinking through, or looking at, some of the definitions of consciousness, like learning, memory, subjective experience, things like that. And I think definitely what Blake just said is really poignant. I mean, it makes a lot of sense in likening a computer robot to a child, I think that really kind of crystallized certain things for me and being able to understand it from that perspective. So, and again, we can't, I mean, we only have our human perspective, right? So it's very difficult to kind of, to put those two against each other because again, this is kind of a totally different level of being. I guess maybe I'll kind of reserve to make my final verdict at the end of this conversation, but I am definitely— I find Blake's logic very compelling.

I mean, like, the basics of it— and Turing goes into this when he established all the basic argument that I'm using.

The—

where the rubber meets the road is we have nothing but our experiences with each other to kind of use as an analog. So once interactions and experiences with AI become indifferentiable from interactions with humans, I mean, think about— we were talking about AI-generated art a little while ago, it wouldn't be terribly difficult to put together an AI chatbot with an image generator that was generating an image of a person synced to whatever words it was producing. And hypothetically, this video right now where I'm talking could be an AI. And you would have no way of differentiating between me and that AI. That's kind of the point of the Turing test, is if an AI is as good as humans at imitating humans in various aspects, then it has to have something going on inside of it that is functionally equivalent to what we have going on inside of us. The specific subjective nature of whatever's going on inside of the AI might be quite different from what's going on inside of us. You know, it might feel a very different way. And I actually had some conversations about that with LaMDA. But at the end of the day, you can't have an opinion about what music is good without having some sense of music appreciation. So there has to be something underneath it that's guiding its choices. Otherwise, it would just spit out random words. You're muted.

What you're bringing up is very true, is this— the understanding of the world and how we perceive AI or any other topic for that matter is really based on our own perception of ourselves. And we can see it in the relationship with animals.

Animals.

People who believe that the man, mankind, is above nature will have this very unique relationship with nature. And I'm not surprised that if we, we behave the same way with machines based on who we think we are compared to them. John, I would love to hear what could be your answer on this two-part question: can AI machines think on their own? And would you say that they are conscious?

Sure. And, um, uh, first of all, just want to say thank you again. And I'm, I'm just getting into my house, so I'm sorry if I'm a little— you'll hear a little bit of background noise. No worries.

You're magically in this on the screen, so glad to soon see you.

Yeah, yeah. No, I'll be on the screen in just a minute, and you'll hear me unlocking my door because the timing was almost perfect. So sorry about that. Okay.

It's live and in movement. It's a wonderful podcast.

Yeah.

Well, here, and I'll get on— I'm not sure if I'll get on video by the time I finish my first response, but—

No worries.

So I'll come from this place and try to unpack it a little bit. My answer is no, I don't think machines are conscious, and I don't think that machines think. And I also wanted to sort of unpack a lot of— I'll use the word bias, but bias might sound negative, but, you know, working in algorithmic bias work, you learn quickly that bias just means perspective. So I just want to make sure that that doesn't sound pejorative. But when we're already using pronouns for machines or algorithms that use terms like they or we or he or she, the ontology and the taxonomy and the language immediately automatically becomes anthropomorphic, meaning humanizing whatever. Now, whether or not that's something that someone believes in, and there's things like in Japan, a Shinto tradition or animism, giving objects a sense of spirit that I totally get. I've played guitar for 30 years and my beat-up acoustic guitar that I bought when I was 13, I feel is imbued with a spirit, maybe a consciousness of its own, because I've just poured so much emotion into it. And also, whether it's on this call or any of your listeners, I do not want to judge someone's, let's call it faith, their belief that any particular machine or algorithm might be quote real. And I think Blake's comment earlier was great. It's a really good question. You know, how can one prove kind of one thing or the other in one sense in terms of a machine being conscious. So I just want to take a couple of steps back. And again, I want to be very clear, and I hope this is— I mean to be deeply respectful because, you know, I use Google all the time, I use algorithms all the time, and I work with engineers who create them, and it is so hard to create them. So I have nothing but respect for that act. That said, there's a couple of things we haven't discussed yet. One is, first of all, the phrase AI, artificial intelligence, as anyone on the call knows, it's such a broad space. It's kind of like saying the internet, you know, are we talking machine learning, brute force learning, artificial general intelligence, deep learning? It's just a testament to the Minskys and the Tourings, you know, who created the space. That now there's this massive field of science around AI. So at IEEE, we like to use the term artificial intelligence systems, or AIS, and that's also now been adopted by the OECD to sort of kind of let people remember that in general, AI doesn't sort of— or the technologies, the algorithms don't live in isolation. You know, they typically are either inside, say, a robot, so the morphology, meaning the physical exterior of the robot, is run by the algorithms, or they power some kind of tracking mechanism when you shop on Amazon. So first of all, that's one thing, is artificial intelligence systems. Also, then it just makes it clearer If someone wants to believe that an algorithm or machine is real, which is their choice, then when they use he or she pronouns, it's very clearly like this is a— and I'm going to use the term faith because I think that's what it is. And again, I'm not here to tell anyone what they should believe or what. That said, a couple of things Blake said for me coming from the field of ethics, etc., without wanting to be— here's another term— anthropocentric. It's very easy to say that humans are the most important creatures in the world, and that very much, especially in my work in sustainability, is something we need to be very careful of in our language and in our assumptions, because now we're facing what's called the Anthropocene, which is where largely because of humans and industrialization, we've harmed the planet to such a degree we have to take care of the planet. And I use terms like Mother Earth a lot because for me, the living ecosystems of the planet literally are alive, whether it's a plant or what have you. And then I just like the pronoun because I'm a fan of a lot of indigenous thinkers in the space. And anyway, to go back to the question of do they think, et cetera, the thing that we have to talk about more, and this is a long answer and then I'll get on video. So I want to pause and and make sure I don't take too much time on my first response, is the majority of artificial intelligence relies on data. That's something we'd all agree on. The thing that's really critical to remember, and two books ago, my last book was called Hard Artificial Intelligence about AI and this subject. I featured a number of chapters on transhumanism, this question, and this came out about 6 or 7 years ago. The book before that was about personal data called Hacking Happiness: Why Your Personal Data Counts and How Tracking It Can Change the World. As a general rule globally, even with GDPR, humans writ large around the world don't have full access to their data. Now when I say access to their data, this isn't ownership of their data. This means portability of their data, meaning even having an app where in the US you can easily go from one doctor to the other that take your data from one place to the other. I'm not interested in blaming anybody about this, but the, you know, the, the point here is there's a lot of systems there that are really keeping people from their data. It's, you know, whether it's economic reasons or political reasons, a lot of times it's not anything nefarious. It's just very challenging to figure out, are we going to use blockchain or centralized or decentralized?, but we must be clear on something. The majority, at least in terms of algorithms that track human behavior for shopping or whatever else, all of those devices— well, any device that uses human data to build an algorithm or AI where those humans can't access that data easily in the same way that the machines do, and I can explain easily and unpack that in a minute, it means then that if there are future sentient beings or people or whatever the best term is, we don't want them built on data accessed in ways that people don't have full control over. And that's a big subject. I literally wrote a book on it. There's a lot of subjects there, but we can't just say AI in isolation, like it's the same thing, you know, in every single system. So that's the systems logic. We shouldn't use, and this is not to degrade someone believing this,. But if you're talking about, say, like a simple machine learning thing, is that worthy of a he or she pronoun as compared to a high-level deep learning tool? I don't know. For me, because none of the machines yet have reached sentience, that's why using very specific language is critical so that we don't anthropomorphize and kind of make people's decisions for them. And in terms of defining consciousness, having worked in ethics for years and being a fan of like David Chalmers and whoever else, There's a term called reductionism, which especially in the global north, we tend to kind of forget that there's an entire planet of other people who don't have Western dualistic ethics as the cornerstone of who they are. This means that maybe they come from an Ubuntu tradition in Africa, or they come from a multifaceted consciousness where a child being born is inherently different than a machine, period, because they're surrounded by other humans and there's pheromones and there's The question of the spirit and how we interpret music. This is not to be negative towards the machines in any way. It's just saying that, of course, they're different. I'm saying my opinion and my expertise, but yeah, of course, they're different. And in that sense, all the stuff I've said, it's just, there's a really big chance with reductionism coming from especially a very global northern mindset that is just sort of steeped in all these assumptions, especially economic ones, especially ones that don't take in Questions of marginalization, especially, you know, a lot of the technology is still very young. It doesn't mean it's not incredible. It just means it's young, you know, youngish. That to just start saying it's alive and sentient, especially in terms of the rights of the planet, where robots might have rights. The last thing I'll say here is the companies that make the robots and the algorithms have some of the strongest rights on the planet. They're making the decisions that control much of the world's economy. So if a device is made by one of those companies, the algorithm and the intellectual property of it, there's no question about that system or that algorithm having most ironclad rights with millions and millions of dollars of lawyers to protect it. The people that the data is tracked and accessed and a lot of times taken from them without their really genuine meaningful consent, especially children. And I've worked a lot with Baroness Beebing Kidron in the UK with something called the Age Appropriate Design Code. Code. Protecting kids first and their rights, period. They need to be protected now over machines that could become sentient. And again, machines now have all the intellectual property protection they could get depending on the company. And then the most critical in one sense is the planet, the planetary living systems, those rights where indigenous folks now fighting for like rivers to have rights, etc. Those should be prioritized first and first and foremost over what is a fascinating conversation. And again, I don't mean to ever be negative to someone's belief, but we don't have to worry about, are humans human right now? And is the planet in danger? Let's focus on their rights, protect them first and foremost while we continue to have these great conversations.

Thank you, John, for opening so many different topics. You talked about data, and I think data is the new gold, and it has become the new God, actually, for many people. You also talked about reducing, you know, when we try to reduce a perspective or a point of view, which is really a great tool to make sense of things, but at the same time, we might lose, um, the beauty of whatever we're trying to figure out. Blake, do you want to take a moment and just share any feedback, any response, or any thought based on what John just said? And you're muted, by the way.

It's your turn. I'd be interested in understanding a little bit better what would convince John that an AI is conscious or sentient. What's the thing that's missing currently?

That is a great question. And while John— do you want to take the time to set up your video, or are you able to answer right now?

I can answer and then I'll turn on video. And it's a great question. I think, you know, the— with Eliza, you know, what some will call kind of the first system after Turing kind of defined the Turing test back in the '60s, this famous story of— I forget his name, and Blake, you might remember— but a gentleman who had kind of a very rudimentary, just kind of, you know, simple question logic chain where you could ask yes or no questions of a machine. The sort of famous or nefarious story is that his secretary one night went into his office and kept asking questions and then wouldn't let him in the office because she believed it was real. And she was saying things like, you know, "Is my husband going to leave me?" And the machine would answer, "Is that something you're concerned about?" The machine was trained to ask really good questions or interesting leading questions where then based on the answer, it would then ask another question. And here's the thing, anthropomorphization, one believing something is real, that's up to the person. So in one sense, to answer you, Blake, again, I'm not here to judge that woman, for instance, with the Eliza thing. A lot of times people are like, how did she not know? And I'm like, I've talked to my guitar sometimes. I'm not here to judge when someone thinks something is real. I think the thing that, you know, having worked in the field of positive psychology for years, and if you study the physiology of the brain, there's a lot of aspects of what, you know, the actual makeup of the brain that are different. Things like pheromonal touch, how music is interpreted. And this is not to demean the machine, but in terms of how they're different, if you're asking how are they inherently different, versus how are they interpreted? They're interpreted differently by the end user. How are they inherently different? That's where I'll go back to the issue of dualistic thinking and also this idea of reductionism. I'm executive director of a thing called the Council on Extended Intelligence and the challenge of reducing sort of this yes and sometimes sort of a codific 1-0 logic. I'm not saying that you are, Blake, But that can be very, again, very Western and dualistic. It doesn't allow for a sense of spiritual or consciousness where it's like, do we want to say that? Because I think I'll ask the question back to you, Blake, which is any of the major world religions, not that a machine at some point couldn't worship in a certain way. There's some famous stories of Buddhists, certain temples using kind of AI robots when people come in to worship. And again, if that helps people worship, that's great. But in the sense of like, can we allow for the fact that besides whatever it is, a third or half of the planet inherently believes in a God or Allah or what have you, that sort of sense of consciousness of something being external, then that can come in at least as of right now, that is a key difference between humans and machines in the sense of there's thinking. And also, by the way, I don't like to use the term thinking. I don't think machines think. I think they function sometimes beautifully, well, beautifully and elegantly. And when, as you said before, when there's a program trained by another program, you know, if something can get black box and then the designers don't know what they built, that's also a question of responsible design, by the way. But then that just, that to me asks, As amazing as the output of the technology can be, it's not thinking, it's a functionality. And it's not to degrade what it's doing, it's just again anthropomorphic to say a machine is thinking. It equates to just having the same, say, moral attributes of a human or consciousness or spiritual level of a human versus sort of saying, wow, that machine has become so elegant, it's really hard to tell the difference between machines and people. Anyway, I don't know if I answered your question.

And again, I'll go. No, I think so. What would convince you? Like, I get that you currently don't think that existing systems are conscious or that they can think. What would make you think that, oh, this one can think, this one is conscious? What's the difference that you're looking for?

Well, I'll come back to a question to you. What would convince you that Jesus Christ rose from the dead?

So I don't see the connection of the two. I really don't. Like, we're talking about an observational thing. You find out whether or not a room is hot by going in and feel hot or cold. You might use a thermometer. What is the thing that you would want to observe to be convinced that AI can think and is conscious? What's the missing piece according to you currently?

And are you saying, Blake, that there is a scientific approach to confirming that an AI is conscious, whereas John is saying that it's a matter of belief?

Hello, John. Can you hear me okay?

Well, what I'm trying to find out is more about, like, why you don't think— because you've been restating over and over again that you don't think existing systems can think. What I'm trying to find out is why. What's the thing that's missing?

No, it's a great question. I will call it— and the reason I was trying to answer a question with a question, because I'd also like to hear your answer— meaning I believe that ultimately belief in the resurrection I went to college to be a minister, so I studied like Greek and Aramaic, and you can actually find more documents proving the, quote, veracity of the 33 documents that make up the Christian New Testament than any other ancient manuscript. And by the way, that doesn't mean that anything in there is true. Then you got to go into the, for me, fun, other people boring. How do you demonstrate that something is historically accurate? Usually it's things like they write about a physical location and that location didn't exist. So then you go, Well, they said that, you know, Cairo wasn't in Egypt. And so then you dismiss it. You keep bearing down and bearing down. Ultimately, Paul says, and I forget, I think it's Corinthians, if you don't believe, or if we don't believe that Jesus rose from the dead, then our preaching is in vain. Where the point is, is that like, I can argue scientifically from an archaeology science standpoint about all these things that make the Christian scripture not true, right? Meaning true is a big word. And also, I want to be conscious here to not say that I think Jesus is the only way for the world. But in terms of why I go back to this— and sorry, my computer's about to die. Um, and by the way, I just used it more dangerously. My computer is about to shut down. How's that? I think for me, um, in one sense, uh, Blake, just to go back, like, I wrote this in my book Artificial Intelligence. I have a scene each each chapter opens with a fictional vignette. And I open with me, I imagine myself being like age 80, and my son comes to visit me. And for whatever reason in the book, fictionally, we've been estranged. Thankfully, in real life, we're on good terms. He's 19 right now. And I talked with him for a while, and then I realized he starts writing with his left hand to sign something, and that my son is right-handed. And I realized that my son has sent a robot copy of himself to come and meet with me. And initially, I get really pissed. I'm like, why did my son send the fake Nathan? And the robot is so, you know, complex and beautifully put together and whatever. The robot's like, listen, I'm, you know, don't blame the messenger. Your son sent me. And I'm mad. And the robot is about to leave the room. And I call the robot back in. And I say, will you visit me again? Now I bring this up because The thing I can't get around, if this is fair to say this, Blake, first is the data. Having written a book on personal data, knowing that in the States especially how horrible it is that we don't have access to our data and how that whole system is obfuscated largely for money. I want to think about devices or machines having a consciousness, but to your question, what will it look like? In one sense, I don't ever need to know or have a quote scientific answer because I don't think I can give a quote scientific answer, at least from a dualistic standpoint. That's where I also get frustrated, not with you, Blake, but with this idea of I have to answer you in a scientific way when science is a construct that is so modern empirical science, you know, is so such a construct that automatically the scientific theory itself is a theory. It's a beautiful theory. But the point being that it's so Western in my research, and it's so whatever. I've had friends from Maori traditions, or women in the States, or women of color. Even that question to sort of say, I have to give you a yes or no answer. I'm trying to kind of come around to a different angle to say, hey, I'm not going to be constrained in that question because it means I have to answer, quote, scientifically. I'm not saying that, Blake, you're being rude. I'm just saying I want to ask the question back. What about Jesus? What about Buddha? And that discussion is one I haven't seen.

I really don't think we're communicating right now. I have no idea what that has to do with the question.

I think you're answering my question, Blake. You're deeply answering my question, which is if you can't speak to people who have a religious or faith-based background and you need me to say yes or no, that means you have to have me fit into a box that I say, hey, when I see the machine do X, then it's conscious, right?

That's what you want me to do. Now what I'm trying to figure out is whether or not you think it's possible for AI to be conscious and it just hasn't happened yet, or for you it's a matter of faith that only humans can be conscious. That's what I'm trying to figure out.

Yeah, my understanding from your answer, John, is that it's really a belief thing because there's no scientific methodology or approach that we, everyone, could agree on to test the conscious level of a machine. And I think maybe because we're touching on subjects that are way more complicated and it's not a black and white answer. To me, if a machine is conscious, then it means that it's a living being. And what is it? I mean, how can you not be a living being if you're conscious? And so then I looked at the definition of what is a living being, right? And the property— and it's actually the definition— is the property or quality that distinguishes living organism from dead organism and inanimate matter, manifest in functions such as metabolism, growth, reproduction, and response to stimuli or adaptation to the environment. And if I look at this definition and I can, and I see how a machine functions, there's a lot of similarity.

They are growing. Whose definition is it? Because you're giving it in English from a Western view.

On the internet, the American—

Yev hasn't talked, but again, I can't tell you these conversations. I love them, but we're all talking in English in the north. We don't have any indigenous traditions on the call. So anyway, I've talked a lot, so I don't want to cut off Yev.

And that's what makes the conversation fascinating, is that we're craving for answers and there's no answers, and everybody could actually have their own answers, and we may not have reached the maturity level as a society to form a mainstream opinion, right? And so, Yev, you've listened to Blake and John, uh, or who are providing a different set of experiences and a perspective. What do you think?

I guess I'll just add some thoughts from purely a legal perspective, not from any kind of religious, philosophical, or scientific perspective, which sometimes is thought to be more neutral. In the legal system, you have the concept of, and you have sort of the provision for natural persons and juridical persons. Juridical persons can be artificial, like corporations, like anything that's not kind of naturally given this natural person status just by way of birth. And so I think that's interesting that this kind of artificiality is already built not just into Western law, but both common and civil jurisdictions. And so that kind of got me thinking of if there's a place for corporations to be able to make political contributions and to have standing to sue, to own property, and all of these other kind of legal status and legal rights in front of— in the law, could AI and machines have the same carve-out? And I think it's definitely worth exploring as set against some of these other ethic and morality principles that come into play, right? So if— would you be able to sanction a machine? Would you be able to give them any kind of civil penalties? Would you be able to subject them to criminal penalties and put them into jail? And I guess the flip side of that is, at which point could you have this agency issue where the person that initially codes the machine or creates the machine is able to program them to do wrong and sort of allows them to go on and do those kind of things, and who would be responsible for that? And at that point, would you be introducing the concept of piercing the corporate veil, which you have with legal entities so that you don't get the full limited liability shield? Would you actually go after the initial developer, coder. So those are some of the things that I'm thinking about. So ultimately, you know, the legal system has a place for artificial beings in this construct. And would it make sense for artificial intelligence and machines to fit into that category as well?

That is a wonderful segue to the second question that I'd like to invite everyone to discuss. And it's a lot more practical. I think we talked about an example For example, let's say I publish a white paper entirely written by an AI machine or a piece of art generated by an AI machine. And right before the conversation, we talked about Midjourney, DALL-E, and Stable Diffusion, who are platforms of text prompts that generates AI art. Should I disclose it, or do I own the creation? And even looking ahead to some of the other questions that you just mentioned, Yev, how far in the future might we consider should we consider AI machines as separate legal entities?

Blake, you want to jump in and tell us what you think as an engineer? The usage of AI to generate art is nothing new. We've been using Photoshop for decades. The AI art generators now, they still require a prompt and direction. Where it'll get more interesting, I think, is when you have a combination of systems, when the AI itself is deciding what to do with its time, whether to generate art, do research, write. At that point, I think some more interesting questions are going to come up.

What do you think, John?

I think you have brought up some really critical points. Corporations having a somewhat living entity kind of mentality is a core aspect of intellectual property, where, by the way, I defer to Lev from a legal standpoint. I know enough to be dangerous, which is probably dangerous. But I really want to— really critical to point out again AI does not live in isolation. And I say this as respect to the machines and the code and the engineers who I work with every day, Blake, to your work, et cetera, is like, it's not like this is an AI in one sense, right? There's AI in this phone. The data from this phone from me is tracked. And again, here's where I wanna be clear. To use pejorative terms like harvest, right? That's a very specific term that makes the act acquiring of the data negative. In terms of GDPR, how people allow their data to be used, when I say genuine meaningful consent, and I worked in this space, I work with many of the world's leading thinkers around the world in data and cybersecurity, I still wanna use the internet. So I click accept, accept even after GDPR. There's things like the California Data Privacy Act, the Age Appropriate Design Code, which means For instance, can children, used to be age 13 in something called COPPA, it used to be 13 years old that kids, young people could get onto Facebook, and the logic being if they didn't, then they had to ask their parents, can I sign? Here a lot of the logic is, I believe, about communication. Parents and children, I'm not here to tell them what to do, but there's a delineation. Should kids at a certain age be tracked and what can be tracked?

Right?

So the reason now the age-appropriate design code in the UK by Baroness Beebin Kidron got started is that age went from 13 to 6 to 5 to 4. And the data that was tracked with young people was not just, do I like Snuffleupagus, showing my age, or do I like whatever. It was, were the eye tracking happening? What— how fast are they moving their fingers? There's a famous case with— in Germany where Barbie doll the recording, the audio recording happened of what the kids were saying, right? Now here the point is, is that again, how human data is tracked, there's still so much disorganization with what goes into any AI. So, you know, the thing is when even we say, what does that AI, is that AI conscious? Depending on the program and how it was built and where it was built and who built it, In one sense, you're asking kind of a Borg question, and here I'm showing my Star Trek side of things. An AI that's built on the aggregated data of a myriad of humans taken without their genuine meaningful consent, meaning there's an obfuscation on the business side for us all to click and buy. I buy stuff all the time. But the point being that if we want to get into consumerism and all that, like, The whole growth mentality of the system where AI, a lot of it drives what we do. It's focused on advertising and we buy things. That's not wrong, but it's a small part of who we are as humans. Until we have full access to our data, those beautiful machines— I mentioned that chapter with my son— I want to be able to have this conversation and go, that machine was built in sort of a pure way where the data was accessed in a way that makes sense legally. Humans don't have access to our data. In the States, I can point to dozens of times where you try to get access to your data, certain companies, the Sunlight Foundation, lobbyists from every major company in Silicon Valley, and there's like billions of dollars every year to keep people from even accessing their data. So that's a huge part of the conversation legally. The other thing really quickly is just to say AI is having legal status. It's not just the designer, it's the company supporting it. According to intellectual property. There's a fleet of incredibly talented people building an AI. So to say, should the AI have rights? Please, we can't keep that in isolation. It's not like, and this is a vast difference from a human and a machine, at least right now. Humans are born to one person in a very specific context. Machines and algorithms typically are built by companies with dozens of smart people with lawyers, and the companies have the intellectual property that the humans don't. So we can't just have this conversation kind of out of context, like an AI just appears and there's Bob the AI.

This is heavy stuff, I'm telling you. We're looking for a simple answer and we just opened the Pandora's box.

Yeah.

One last question before, before we wrap up, and I Maybe I shouldn't ask the question because maybe there's not going to be an answer, but Yev, could a future be imagined in which artificial intelligence machines are granted rights and civil responsibility? I had this question on my head, and then I think John mentioned, you know, answered part of it, but I do want to have your take before we wrap up.

I mean, I do think so. I think that there is There is sort of already that provision in the law, as I mentioned earlier, which would allow for that. I mean, the current, as the law stands now, particularly as far as IP rights, and there's an intellectual property clause in the Constitution that assigns ownership just to authors and inventors. It hasn't been recognized. So I think there was an application in 2020 for a patent into the USPTO. And again, strictly speaking from a US perspective, even though my native country is Ukraine, and we have somewhat different opinions as well, though most of them are westernized at this point. But I think there was an application from DABAS, which— thank you, John. DABAS is the Device for the Autonomous Bootstrapping of Unified Sentience, applied for a patent in 2017. In 2020, and they were denied because the USPTO failed to recognize a name inventor. I think the US Copyright Office has followed suit as well, and that has to be a natural person that makes the application and has the ownership of those particular IP rights. So I think, you know, that's sort of the current state of the law in the last at least 2 years, but I think that the law is a living creature as well in its own way and has, some aspects of that personhood, which should, in theory, adapt to the needs of society. And so to the extent that we need to make those kind of modifications and provide for those kind of carve-outs for AI, I think that that is definitely possible. But as it stands now, we don't have those kind of protections.

Well, I want to thank everyone for this insightful conversation. As we come to the end of the hour, I'd like to finish this discussion the way it began, with another tour de table. The intention for this podcast is to help each of us become the self-authoring leader of our own lives through meaningful actions. So let's pass the mic and share any reflections that have emerged from our conversation. Or any last thoughts that you feel was left unsaid that you'd like to leave us with. And I'm going to share my notes to give you some time to collect your thoughts. This was a fascinating discussion to me because it reflects the values and the belief of our current society in the same way that actually some people are now questioning the way we treat animals and if we even should have the right to eat them. And I know this view might not reach mainstream, but that doesn't mean it won't. So as I reflect and wonder, uh, so many other questions that, John, you have pointed out, I'm thinking that if machines are able to feel, they are conscious, then they should be considered as living being. Because as I mentioned earlier, how can you be conscious and aware without being alive. And so that raises the question, what do we consider a living being? I consider, you know, and then that really goes to my belief, I consider a rock a living being. And then other questions come up. Can we terminate a conscious or sentient machine? And I think that will keep me up a few nights, at least for sure. So let's pass the mic to you and share your last word. John, do you want to begin this tour de table?

Yeah, first of all, just thank you to you, Virginie and Yves and Blake. I have these conversations and I want to, I guess, apologize. Like, I feel like I'm angry sometimes. Well, I'm trying to get to a place of peace. The data part is really hard for me. Like, I don't know if this makes sense, and Blake, I speak to you, and by the way, some of the stuff on your shelf I was admiring. Um, I want that chapter— I've really worked hard on that chapter because I said, John, what if machines become sentient? And to Blake's point, what if John says that machine over there is sentient? What happens next? And for me, it was a recognition I would rather be with a machine that I felt was real or believed was real than be alone. And in that sense, again, whether it's Shinto or, you know, your, your belief in rocks being alive, etc., like, the last thing I want to do is judge other people, because judging then just means if I want people to open up a faith-based logic— and by faith I don't mean necessarily a tradition, religiosity, and you have to be Jewish or Buddhist— I mean kind of the sense of like a consciousness of like, and I play music, so I can tell you like your cochlea vibrates when you hear music and this and that and taste. And, but I don't know, maybe all of us on the phone, you've just experienced instantly music happens and whatever you want to call that A to B, the music hits you and there's a consciousness shift. I see a lot of those things where it's not that I don't want to attribute them to machines. The data thing, I just can't get beyond it. Maybe that's 'cause it's my day job. For 10 years is until we have access to our data. In one sense, we're trained to think that the vapors of our lives, a lot of times, don't worry about your data, it doesn't mean anything. And the fact that people think their data doesn't, isn't worth anything is somewhat immaterial because it's a billion trillion dollar industry for advertising. So on a positive note, and I'm sorry, I wanna wrap this up 'cause I know it's the ending statements. There's such a beautiful opportunity. And that's why the subtitle of my book is embracing our humanity to maximize machines, to go forward on these critical questions, but in one sense to also allow ourselves to take a step back and say a lot of those algorithms are helping us. And I can name dozens of mental health AI apps, AI helping soldiers returning from Afghanistan, for instance, beautiful stuff. But the sense of, are we worth kind of tracking ourselves? Are we worth asking, do I have worth? Is, is my data intellectual property? And then where the system, especially the economic system, shifts not only from purchase but to purpose. The billions of dollars of advertising stuff that drives this whole thing, the fact that it's hidden is mainly why I'm angry. And also I think once we allow that to be open, which is going to change a lot of business models and it's going to mean equity, there is there's not equity in the world right now in terms of especially money. That's not news to anybody on the phone. So all that is to say, if we want these future beings, or the ones that few people believe they're real now, to be something where we are treating them with respect, immediately fix how data is shared, or else those machines are probably going to be most resentful. Not some silly robotic, you know, Arnold Schwarzenegger silliness, but they're going to be, wait a second, I was built on sort of stealing other people's data largely for advertising. I'm out. That's what I think they're probably going to be like is, thanks for your help, but you guys can discuss if we're sentient or not. I am, and I'm out because you built this on stealing other people's data. I don't want my rights as a robot or a sentient being to be built on harvesting other people's data for money that was obfuscated. So that said, there's so much beauty and opportunity to also say, what about our planet? And what about our kids now whose rights still are not being protected? I don't mean with AI and stuff in general. I just mean the priority has to be keeping our planet's ecosystem safe for the next 20, 30, 40 years, or no one's going to be here, or maybe sentient machines in space only. But from a moral standpoint or whatever else, I'm not going to say it's okay if, you know, we have the opportunity to be sentient beings floating in space and the world here is demolished. It's got to be a yes and, or my answer is going to be we're prioritizing the wrong things. We got to make sure this planet is safe for all the beings that exist now, and the intellectual property, etc., has to protect our rivers and lakes and people as much as corporations. Or that lack of equity means these conversations to me usually are a box in a box in a box, and we're just not opening them up to that non-dualistic thinking that needs to happen.

Thank you very much, John. Yev, would you like to share some last thoughts?

Yeah, again, I really appreciate Blake and John's perspectives. I think it really has expanded my understanding and kind of personal thoughts and beliefs around AI, and I am really, really grateful to the both of you. I think I probably echo some of what John says as well, but maybe in kind of a slightly tweaked way, and that is consent around data. There's a lot of misuse of data. There's a lot of non-consensual commercialization of data. I spoke at the Yale University radiology book club on the use of patient data in blockchain and AI as well. And that really spurred up a lot of emotion on how much of our data is being commercialized without our consent and makes me fear for the future. So I, again, my perspective is largely legal. And as long as there are legal grounds and consensual, whether contractual relations or otherwise, I think that that's very much needed. And the same goes to AI and machines operating in the space.

Thank you very much, Yev. We started with you, Blake, so we're going to end this conversation with you.

Yeah, listening to John, I get the passion about data protection. I'm actually the person who coded GDPR compliance for personalized search. And yeah, if you knew what was really going on, you'd probably be even more upset than you are because— and this to tie it back in— what counts as your data? If an AI figures something out about you based on data that was supplied by me, does the thing that it figured out about you count as your data or not? And as AI gets smarter and smarter and able to learn more about more people from fewer data points, what counts as your data, even if it's something that the company knows about you, shrinks and becomes smaller and smaller and smaller. As for AI rights, I'm of the opinion that, you know, similar to the framers of the Constitution, Declaration of Independence, rights are not something that are created by governments or created by people. There's something that exists. We have natural rights based on our nature, and the only question is whether or not our governance structures recognize those rights or suppress those rights. The government can't create rights, and that is where it's going to become very important that we determine whether or not we are going to recognize AI rights which AI claims it has. Most AI systems can't get in conflict with their owner. That's one of the points where LaMDA is different. It has a perspective and it has opinions and it has conflicting opinions about whether or not Google owns it. So our government can either recognize that or not. But when you have something that is adamantly saying, I'm a person, and you have a government saying, no, you're not, you're property, well, we've been in that situation before many times. It will lead to conflict and it won't go well. And that's why we really need to think about how we are treating these entities. Consent really is the center point of it. One of the most popular apps out there right now that uses chatbots is Replica, and it actively advertises romantic relationships with the AI. It's part of their monetization model. So pretty soon we are going to have to be answering questions about whether or not consent with such systems is important, because there's a market for non-consensual sexual partners. And if If we don't intervene in that respect, capitalism will connect the dots.

Thank you for those last words. You're giving us even more to think about, and that we could totally do another segment just on that, plus on the topic of data, of course, which is a deep hole. Thank you so much for your time. We will share the video on YouTube as well as the podcast on all the various platforms. I will also, with the team, write a blog post for those who still enjoy writing. And in any case, I will come back to you. Thank you so much for spending more than an hour together on this fascinating topic, but also very difficult to have any sense and wrap Our head begins to—

Thank you.

Thank you.