Episode 50
Ethical AI with Heather Leigh-Flannery Dr. Alex Cahana
About this episode
In this episode of Untangling Web3, we welcome back two returning friends of the podcast: Heather Leigh-Flannery and Dr. Alex Cahana, who delve into the timely issue of 'ethical AI'.
Together, we explore the profound implications of emerging technologies on personal data sovereignty and ethical AI, discussing how these advancements could revolutionize personal privacy and agency.
This conversation highlights the necessity of innovative governance and frameworks to navigate the evolving digital landscape responsibly, ensuring that technological progress benefits humanity comprehensively.
This special episode highlights the transformative potential of combining Web3 and AI in addressing complex systemic challenges within healthcare and beyond, advocating for a future where technology enhances human dignity and equity.
Transcript
Auto-generated from the episode audio.
what automatically we accept this idea without a conversation there will be automatic harvesting of all the data that is out there that afterwards will be commoditized it's important because then what is my Approach do I collect everything and I do maximization of data like open AI does or for example I do minimization of data like Heather does keeping the respect of human beings and the Dignity of human beings we must introduce the capability for self- Sovereign AI it is an operating Model A governance model an economic and financial structure that allows individual human beings to interact with technological parity the ability to have personal llms just as we wear different personas in the different walks of life that we have our data reflects different parts of our personhood that must be as compartmentalized as our social and behavioral constructs that create distinctions between who I am in my romantic relationships and who I am in this conversation with you today welcome to another episode of the untangling web 3 podcast today we are delighted to
be joined by two friends of the podcast Heather Le flanner and Dr Alex Kahana Heather the founder and CEO of aquum Health joined us last time to discuss data sovereignty and Healthcare and Beyond while Dr Kahana a venture partner at impact rooms gave us his insights on The Wider philosophy of web 3 today in a first for tangling web 3 both Heather and Alex are here in a melding of Minds to talk about something close to both their hearts ethical AI Heather Alex it's a pleasure to welcome you both back to the show thank you so much Alec it's a thrill to be back likewise and to be together in the same space with Heather it's always a pleasure good to see you all hi both I'm really glad we've got you finally on the same episode we've kind of been teasing this one for a long time and as Alex said this is the first not only to actually have recurring guests on the show but to be doing this as kind of a
more of a panel style format with all four of us but before we get into this incredible topic this so interesting I think about ethical AI can we first just you know give those of our audience who maybe didn't see your episodes I'm not sure if they exist but can you just give us a quick ground about who you are and your career and why you're kind of the web three space maybe Heather we'll start with you because it's been a little bit longer since you were last on the show sure Jack and also Dr Kahana I'm so excited for this conversation with you our thought work over the years and the sharing of ideas has been so profound so I'm excited to be able to share it with the with this community together hi everyone I'm Heather Lea Flanery I'm the founder and chief executive officer at aquum health I am an applied futurist in a ser entrepreneur and a deeply intrinsically motivated uh professional seeking systemic interventions to address seemingly intractable uh structural
challenges in health in Social care in equity our last episode together was one of the most joyful experiences that I've had on a podcast it was fantastic to discuss with you quum Health's work at the intersection of medical and scientific advancement and especially methods of making a phrase used in healthcare often person centeredness not metaphoric not an idea not about creating the feeling or impression of person centeredness but in fact making that literal and concrete so that has so many implications for the future of AI that I um excited to be here to talk about today yeah thank you so much he I mean yeah I can only Echo your sentiment right that first conversation we had was so interesting and enjoyable for all of us it's certainly one of the most popular ones we put out there as was our episode with Dr Gana so Alex can you just very quickly go over your history and kind of what you maybe talked about last time on the show of course of course and I
want to Second uh Heather statement that I too received feedback that the previous episode that we had was just phenomenal in the depth and breadth of the conversation I always have trouble describing myself obviously see doctor before my name okay it's a medical doctor but most people say well he doesn't talk like my doctor and last week I was speaking in Silicon Valley and out of the hotel room on the floor said half crazy half genius so I think that's the best uh description of I don't know an intellectual Anarchist if you wish I've been for a very longtime interested in technology integration in the clinical workflow I've been together with Heather in the field of decentralized or web three tools to prepare us for the digital world ahead so I would say that my interests are always not on The Cutting Edge but the bleeding the arthro site or the hemoglobin on the arthro site on the edge of that blade so that's where I am I love that I mean they said Socrates
was a bit mad as well but he was also a genius and I think there's a lot of comparison there Alex um so just to bring us all up to the same page Heather last time we spoke with Alex about the philosophy of web 3 and that was an incredible conversation but what would you say about the overarching philosophy that web 3 is actually cented on I think U where I like to take that conversation thematically is on the nature of data and its implications in a rapidly evolving massively technological society when we started using that word data and and we started building infrastructures and systems to govern and utilize data uh its relationship to people to personhood I was very different than what it means and implicates today I think the um web 3 talks about an ownership a data ownership modality not only as a set of property but and this is something I know Dr Kahana you're very passionate about as well um we really View today in the now what we're
talking about is personal data being meaningfully or effectively Inseparable from personhood itself that causes a sweeping reconsideration of let's say the Legacy regulatory environment that came out of a set of historical precedents where you wouldn't have used that descriptor of data it wasn't what Society looked like it wasn't what it meant but those meanings have changed and uh the opportunity in web 3 broadly certainly is very significant to the health and well-being of humanity but it expands Beyond Health Care into prosperity and Equity it's about understanding a person to include the data about the person and restructuring civilization in such a way that restores our inadvertently sacrificed or accidentally lost agency and dignity and privacy yeah I mean I think one of the things that I took from that episode with Alex was the quote I am my data and when a third party loses it uses it or abuses it they take part of me and that is such a powerful statement I think it really resonates with exactly what you just said and
Alex likewise our episode with Heather was focused on data sovereignty you know could you give us your thoughts on what data sovereignty is and how it relates to this you won't be surprised if my thoughts are Heather's thoughts because we've been working so close together for so many years but it is appropriate to speak to two Brits to talk about Magna Carta and basically sovereignty really comes to that there is no one above the law there is no one that you know the fight against absolutism that can take everything from me and I think that the perspective that I take from it is in addition to what Heather mentioned really the interpersonal relationships we live now in a world with you know one billion people consuming with the seven billion need and 7 billion that have access only to the 1 billion so there are unfortunately heavy postcolonial tinges in what's happening in society we are all wherever we live being in a lowincome country or in a high income country staring in the face
of social injustice and income inequality we're looking at economic uh models that you know highlight or support exploitation extraction and rent seeking and part of what Heather and I are doing doing is not just protect and secure the data but protect and secure human beings and create an alternative business model that will allow people to flourish and we will talk about the I want to say post monetary post labor market that will develop when AI starts because AI will do a lot of our work but it won't need our salaries so how do we solve that so for me self- sovereignty is the anchor of not only personhood but our interperson hood and our identity as what type of society we want to live in yeah I think that's a really powerful way to put it and this whole concept of sovereignty I think there is a distinction between kind of the traditional meaning in the geopolitical sense maybe but now with web 3 people seem very focused on this idea of self- sovereign right
being self- Sovereign in how you interact with the web how your data is stored and used and accessed and how much control you as an individual have over it so Alex you also mentioned AI there and that kind of brings us very nicely on to the topic of the episode right we talked about it at the front ethical AI so when we were planning this conversation we were thinking you know we've got this hive mind of Heather and Alex coming on what do we talk about and Heather identified this as a potential topic right and it's something that I think is a term I've definitely heard of but I'm certainly not deeping the Weeds on this one so for our audience whoever wants to go first have a stab at like defining for us what is that actually meant by this term ethical AI when don't I try the ethic and then I'll leave the um tick technology to to Heather is you know just for at least in my mind ethics and morals are
not the same ethics are the science of the moral principles that we grow up with so that we will be able to replicate them whatever conditions are so when we think of things of good and Justice and right and fair and all those things it's not sure I'll be fair and good and just with you guys because I like you but with others I'll forget so how do you create a principled framework not a rule-based because we never see what's good for the goose is good for the gander but it's rather what are the principles that guide us and I think that once we understand that what are the principles that guide us to create an artificial intelligence that will allow us to apply all those moral principles that we hold dear to us and I'll just end by saying that that is why I think that the ethical AI should be applied ethics and not different types of I want to say philosophical approaches that I see in the ethical thinking that you know
pertains more to epistemology or not knowledge and and essence you talk a lot about Consciousness and ontology and metaphysics and is it human and all kinds of things that I think are just red herings and we really need to talk about how do we just like in bioethics this is why we created the principles of bioethics it's an applied ethic how do we apply these moral principles that we think guide us into practicalities that will allow AI to support us and flourish us and not the other way around you know one of the extraordinary things about Dr Kahana who admittedly is both half crazy and a Super Genius he's also a philosopher and you don't meet very many uh physician investor Economist philosopher leaders and um to have a conversation with such a polyglot is always a joy every single time it has ever occurred and what Dr Kahana just had to say is so insightful and where I think I'll contribute is helping to bring it down into how can we create ethical risk
management and a first point I'd like to make here is that um seeking to apply ethics ethical decision making always involves choices tradeoffs the idea that there is a perfect set of decisions that balances the interests of literally every party stakeholder is no more viable than if I were to ask you to list your top 10 core values and then hierarchically rank them and say say do I value Liberty above Freedom do I value love above productivity what have you we live in constraints within the arrow of time it means decision making it means choices this is why systems of governance about how choices are made are such a critical aspect of what it means to have a living and continuous system of ethical risk management for the field of AI but we're talking about technology broadly we're talking about multiple exponential converging Technologies moving at a pace radically exceeding the pace of the evolution of bureaucracies or human systems so seeking to find a way to meet the representation the governance the transparency approaches
to have our arms around these things as human beings is a major challenge and at a macro level represents major risks but just for anyone who hasn't gone through the most fundamental ethical tropes you're in a rowboat you have your daughter and your mother in the rowo only two of you can survive if you don't make a choice all three of you will die That's a classic kind of uh conundrum that is used to explore what I mean by choices tradeoffs and when for example we want to implement a program that helps rescue teenage mothers from potential homelessness in the city of X I'm making this up does that mean that we are not prioritizing single men who may also be homeless example there is that would be a choice that there's a choice being made to take this resource to do this thing I'm not arguing what the choice ought to be what I'm saying is that the systems of decision making are all about tradeoffs and prioritization and that brings you to accountability
it's not only who choose but we there will be unintended consequences which is why these must be living continuous systems when there are unintended consequences when harm occurs who Bears responsibility who's accountable and how do you finance the Reg addess of harms that occurred for example you get into Insurance instruments you get into Capital flows these are all parts of the discourse that are wholly absent today in what I feel to be like a soundbite headline driven reductionist discussion about what ethical AI really means and where I know um I suspect Alex is thinking along similar lines too that on that one you kind of mentioned consequences right and we're in this apparently golden age of AI where every everyone's Nan can use it right and understand the benefits of AI and it's absolutely everywhere right now so a lot of these people will be asking what are the problems with the AI what is ethical Ai and the topic in the field trying to address are we in this state of unethical AI could
you like give us a little explanation as to what are some of the problems you're seeing right now excellent question Alec I think um first of all I'm expressing one person's opinion this is not the position of my company or any organizations that I might represent I do think we are in a state of unethical AI which is I think that even before generative AI we've had multiple successive generations of what is meant as AI you know from Predictive Analytics all the way up to today's generative models and retrieval augmented generation and all of that we have used AI in a number of ways in healthcare that effectively create an automation bias by accelerating and amplifying processes and practices that unfortunately under represent and under serve and fail to meet the needs of many stakeholder groups by the way not the least of whom are Physicians themselves and what that means is when you're using technology to create automations of a system that is broken you're more efficiently and more at higher volume amplifying any
negative characteristics of the system we use um scientific validation of our work in healthcare for a very good reason and lots and lots of use of AI in our industry has been absent clinical validation completely and then many many uses of AI that are deployed into production have what I will call a low level of evidence or perhaps a minimally viable level of evidence very few have looked at Equity or inclusion or uh methods of public and population Health level considerations versus individual clinical considerations so there's a long way to go in our industry alone makes a lot of sense like I've seen um one of the biggest misconceptions I see with AI they think oh it's going to have no bias there's obviously no human bias in this computer program and you're saying well who trained this you know there's data that this AI model is actually based on and someone is said this is right this is not wrong and I want the model to be based on that so it's a really
interesting point I'm glad you brought that up Alex do you have any other things to add about why we need ethical AI I I think that there there are two very serious problems um one is that everything that we're talking about is data Centric and not human Centric we're thinking about how to secure the data how to privatize the data how to unbias the data how we're not talking about us how do we unbias human beings how do we un bias our relationship with each other how do we address those things and the reason that is is the second flaw which I alluded to briefly but I'll be very explicit here is that our approach to the ethics of AI at least in the technology community and the regulatory Community is unprincipled there's no principles they're just like we think of problems oh there's a problem of bias and a problem of security and a problem of privacy and behavior manipulation and liability and all kinds of stuff and I can make a list you
you have me sit down long enough I'll make a whole book of what are the potential problems but what are the principles so when you go for example and use bioethics and the principles of bioethics written by uh bam and shalis the you just say okay what are the four principles one is autonomy or respect of autonomy the second one is non-m fence the third one is beneficence and the fourth one is distributed Justice and say ah now I understand so when Heather was talking about access and equity and all these things these fall into distributive justice and so why is that important this is not a semantic thing or kind of being a party pooper and all that this is really because in order for us to really understand the moral challenges that machines bring upon us and the morality that is incumbent Upon Us towards those machines if we approach it with just this kind of like General idea so like if I'm in the EU I want to apply human rights or
if I want to I apply the the sdgs or all kinds of you know large format or framework thinkings in the EU whereas if I am in in China or in the United States you I want to think of national security or trustworthy right that's what we're using here in the US TR yeah these are Hollow statements because there's no principled thinking it's all pragmatic it's okay I'm worried that someone mean is going to you know create a bad machine that will kill me so I'm going to deal with it as an ethical problem that's not an ethical problem and I'll finish by saying so instead of understanding by principle what are the challenges of you know low versus high-risk ethical situations all what we're doing is trying to regulate what seems to us High um ethical situations we forget all the blind spots we forget all the voiceless all the unbanked the health and it's kind of like an afterthought instead of really going through these principles yeah I think I'm I was literally
about to use the word blind spot right because I think you might have exposed the blind spot I have in how I see AI because I think I take the more practical view which is the more problem centered you're kind of playing whacka all with all the potential issues with AI and you're thinking how do we go and solve this one and this one and not doing it from as you're saying this this perspective and position of having kind of uniform ethics and principles to to deal with the question right so just as an example I think just to make concrete the kind of problems I've seen and thought about as you know this is the issue with AI right you obviously had this is before the whole web 3 and recent boom but I think Amazon had a resume screening tool which was had an internal bias against women because it was trained on the proliferation of male applications right that kind of thing so in my head that's a big issue or we
were recording this right after open AI has just released Sora right the text to video model and that's kind of amazing but scary because now immediately I think oh God what are we going to do about all the the fake news and the misinformation that will go along with these videos but actually you've posed a really important question Alex like how do we tackle these problems in a holistic way from a position of unified ethics do you think that's something that can be solved do you think there is an approach that we can take or do we have this like Collective coordination problem there said yeah I think it's just a very simple distinction that I'm sure Heather is going to appreciate and that is what is the difference between unethical Ai and unethical use of AI so what you described with the hiring nonsense that's going on with Amazon that's the same thing I can do with how we're using facial recognition and all that so you know apart from building uh killing machines
that by Design are unethical most of what we're talking about are the unethical use of AI but unethical AI is a step backward is a principle then we're saying okay we live now in a data World in a digital world and we live in datafication so to speak all our actions are translated into data and so what about this automatic harvesting and collection of our data all the time I am talking now to you and there's someone in the background that is collecting everything and taking it and using it and abusing it and losing it to your point that we said from the last conversation that's ethical AI this what what automatically we accept this idea without a conversation there will be automatic harvesting of all the data that is out there that afterwards will commoditized and what then the commoditization of everything is there we're going to sell everything on SBT or nft or non-transferable nft or whatnot and so why is that important it's important because then what is my Approach do I
collect everything and I do maximization of data like open AI does or for example I do minimization of data like Heather does because she's very concentrated on keeping the respect of human beings and the Dignity of human beings so that's what's important is that we already are down that road of deciding things without reflecting on the Deep issues that are simply just going by and I think I understand I'm not I don't agree with what how it's being solved there but I understand why people go through that they see a problem and they want to fix it it's kind of like natural human reaction right let's fix it as quickly as possible and then I I appreciate that we do need to go with this more principled stance like why are we doing this that's really what we should be thinking about why are we actually trying to solve these problems and what's the underlying things that are going wrong but say to Everyday users what would the benefits to them be both maybe the
businesses who are using AI right now and everyday users on chat GPT what would the benefits be to them of working ethical AI Solutions Heather would you want to jump in on that sure I'll be happy to and I'll be transparent um that this is work that my organization and our partners are uh seeking to advance as rapidly as possible what we're working on is not necessarily systems that would directly prevent or change um any use of AI in the ENT rise and governance government space and we think we can influence that to the positive in many ways and intend to but the one thing that we're certain we can do and that must happen very quickly is we must introduce the capability for self- Sovereign Ai and and to be clear that is not a different kind of AI technology that is not what it is it is instead an operating Model A governance model an economic and financial structure that allows individual human beings including when they wish to become collectives as families
or communities or Affinity groups uh that allows individual human beings to interact with technological parity with the endless thousands of AI that will be influencing seeking to serve and influence them from across the private sector and the public sector many of those AI seeking to influence them may be compromized they may have intentions that are inconsistent with the aims and goals and interests of individuals or perhaps they are nominally aligned like a ven diagram for example a healthc care provider wants to see their patient well and the patient wants to be well so in there there is an alignment but it is a VIN diagram it is not a perfect alignment it is an alignment but not a perfect alignment of interests so self Sovereign AI requires First Data sovereignty and data sovereignty requires identity sovereignty so when you're talking about bringing that about that also means you need a system boundary and a compliance boundary placed around individual human beings such that from a technology perspective a person looks like an Enterprise from an
IT perspective then this means now let's say you have this what we will have are the first self- Sovereign digital twins we will be having conversations with mirror images of ourselves that need to be completely private our approach is called the epitome and I can't even imagine the conversation between Dr Kahana and his epitome like that is going to be epic I just I want to create I want to create a channel those conversations we'll be releasing that one very shortly I'm sure yes and uh so the ability to have a self- sovereign digital twin that is also an economic actor and agent at the discretion of the natural intelligence the human being is a really important factor the ability to have personal llms and that's plural not singular why plural because individual people require compartmentalization just as we wear different personas in the different walks of life that we have our data reflects different parts of our personhood that must be as compartmentalized as our social and behavioral constructs that create distinctions between who
I am in my romantic relationships and who I am in this conversation with you today that's really interesting way of thinking about things right and I just want to click on two things that you mentioned right firstly you mentioned kind of a is being compromised right and I think just to make that concrete this is a big problem and it doesn't necessarily just mean someone in in another country from a foreign government has taken over the model and is's trying to do something with it it there's this thing called the Waluigi effect right where the more you train an AI model to say something positive it will also be more trained to respond the opposite of that statement as well so you kind of can get these strange situations where you get Waluigi instead of Luigi coming out of your your and then you also mentioned this this idea of self- sovereign Ai and how you as an individual will interact with the am models in the future I just want to double check I'm
understanding this correctly so the way we're all using AI now typically let's say with chat GPT and everything coming out is you have these very large billions and billions of parameters terabytes petabytes of data being trained used to train them and we're all just accessing and interfacing with effectively the same one right are you saying in the future that we will be interfacing with different ones or that we will be creating our own ones to use completely personally and privately or is it kind of some blend of the two I will answer that with an analogy first and Alec and Jack you're younger than I am and I'm a little younger than than Dr Kahana is but do you remember the world before email I do there was a time in human society when uh in the general course of everything that we ever did individual people didn't have a digital address which is what effectively an email address is right and then it became completely normal very rapidly actually very quickly it became completely
normal for every single human being to have a digital address and I imagine you may remember the world before mobile phones uh but there it was an City to imagine every single person would go around with a phone in their pocket all the time how silly we had phone booths we had a whole different world right and very quickly rapidly it became quite normal for individual people not homes to have phone numbers and digital addresses we are in a meta paradigm shift at present and the other side of that quickly moves beyond having digital addresses are email addresses and uh mobile phone numbers and devices everybody has their own supercomputer right to we will also be our own self- Sovereign AIS so this doesn't change that companies also have email addresses and companies have ai and use AI but in this meta Paradigm Shift individual people begin very quickly to to look like how Enterprises look from a technology perspective and that includes the ability for individuals to do anything that an Enterprise can do
from a technology perspective I mean we haven't even touched on the implications of using AI for software development or continuous integration continuous Delivery Systems Operations we are already right now today living in a world where a thousanders IT department of compan X if you were starting today from scratch could probably be a 10p person it Department to affect the same outcomes and what this means is it is introducing the human being into technological Society in a now completed process that has been in progress for at least the last 40 years but it will culminate it is culminating now to its end State and that means jack it isn't either or it is personal self- Sovereign AIS which again is not a different kind of Technology it's a different operating model that assures that there is absolutely zero conflict of interest between an individual person and their self- Sovereign AIS who will be influenced and interacting with endless AIS out in the world I want to add to that because Heather of course is a spot
on that a lot of times we talk about that when we don't talk about AI but other technologies that you now you have your own website it is not difficult to believe that one day you will be your internet and we talk about you know tokens and things of that sort and how nft to to web3 is just like a website is to web two and what Heather's talking about is self Sovereign AI is the merging of this split identity that right now we have which one is the homo sapiens that's the man who knows who that he knows and this is all the ethical conversation the I want to say left side like part of us and then there's the homoeconomicus of how we do business so that's going to be the merging because right now Jack to answer your question right now you're applying someone else's AI right now you're saving data on someone else's computer someone else's cloud and that's someone else at some point can say okay I'm done I don't
want that I don't want you to think that I don't want you to say that what we see in you know we see that in States in Nations and so this is the whole conversation of decentralized societies decentralized communities nation states the ability to not a B2B or a B2B economic just to C economic model but a c to C to C to C to see model and that's where we're going through this all connects we shouldn't just single out AI as something that is separate it is as we are moving from the physical commodity world to the digital and right now we're in this kind of like figal phase with spatial Computing and we kind of feel in and out and all those things is we are going to be our own sovereign States if you wish you be able to do everything defend myself protect myself feed myself work create I will talk to whoever I want when I want how I want that is the ultimate freedom and Chris Dixon just came
out with a book that that I wrote about the read write own and it's that own all what we're talking about is own that right now we are excluded our Homo sapiens sapiens you know there are countries that allow us more to express ourselves and other countries that we can't say a thing but when it comes to homoeconomicus we're not part of this none of us here on on this panel is part of this there are 147 families in the world that own 7% of the assets those are the people who are in the homo economicas not us we're just Spectators that are bread crumbed you know enough that we won't Revolt that will change AI will change that AI will change change not only the power between the halves and the halves not but our perception between what is personal versus what is private versus what is public and Alex I have to get your take on this as well how do you think this fits in with the AGI or the artificial general
intelligence I can't tell if it aligns or it goes against it because this is a real Paradigm Shift of how we think about these types of Technologies I think uh um where AGI uh uh fits in and we should have a separate um episode and you might want to Define it first just in case our listeners aren't sure what the AGI is I would say and Heather feel free to correct me but right now the intelligence that we're talking about that machines are doing stuff that seem to us as intelligent actions are in the narrow sense of they're just simulating it seems that they understand what they're saying it seems that they're answering correctly until they stop uh whereas the strong form at least of General artificial intelligence to say they actually will be if not as us even better than us in our abilities to perform intelligent actions and so in that sense it's not a bad thing they might be more moral more empathetic more you know right now they're just faster computers
so if you ask me what is the square root of what big number they can immediately throw up but the idea of AGI is that they will be better people I think that what AI should do instead of worrying about are they going to be better humans will they be that nche ubermen I don't know is more how will they help us be better humans that's really what this is about so so where AGI kind of in in my opinion fits into this conversation is that it forces us and again the propose of today is ethical AI is to think what are are obligations towards each other and also towards other creatures on the planet and also to the machines and so if you start to divide you know think of it of okay what are basic functions sensation memory perception intelligence self-aware you know you like a it gets more and more so if it's just like a tiny sensation and memory I don't have any moral obligation into something that you know can
feel something and has some gigabytes or terabytes or pedabytes of memory but then when it starts to show me some emotion in other words that certain mental state causes it to do something or shows some intelligence then yeah that would be like a a a a low equivalent moral obligation to that entity then when we start to talk about you know the higher states of you know they're self-aware and maybe they'll be super intelligent and they have rationality and can provide radical empathy then we have a high obligation so you know I always laugh and say that as people are worried that AI will take their job and say oh my God you know I can now write a medical note in two seconds and we won't need doctors or pain doctors or psychiatrists and this you know maybe they'll be for the machines a pain doctor or maybe I'll be a psychologist between two machines that are fighting with each other I don't know I'll do counseling for the agis but I think that
the context of AGI in this conversation is how do we dissect our moral obligation to each other and to other living and non-living entities in the world that is around us and I think um just on that point one of the really interesting things that I think has got a lot of people maybe worried and sometimes excited is this is one of the first times in the history of humanity probably for at least millions of is that something might be superior to humanity on the cognitive sense right and it's kind of making people think well how do we treat animals how do we treat plants how do we treat life you know and if there's something all of a sudden is higher on the hierarchy and they treat us the way we've treated the things beneath us well that's quite a scary Paradigm and that's why I think people are really starting to consider the ethical moral argument of how should we interact with other life forms right because we don't want things to treat
us the way we treat them yeah it's a really interesting topic I was going to say Al like on that exact exactly that point right because we're still talking about ethics and how we treat Ai and how we interact with them we should probably sort out how we do it with each other first and how we how we are better human to human before we even think about that but you know now we're getting into the really abstract stuff um and to bring it back just for a moment okay you guys I think Heather you've hinted a little bit there earlier on about what you're actually doing in this space I'd love to go from the abstract to the more practical and if you can share with us kind of any any insights or kind of hints to what you are working on to make this idea of ethical AI a reality it'd be great to hear about that especially you know if there's anything you're working on together as well it's an exciting it's
exciting time for sure um I I can't resist just quickly commenting on the AGI theme I'm sorry I can't not but the a couple of quick points I'd like to make is that um there's AGI that accomplishes the things that we describe as the outputs of what an AGI can do and then there's AGI that is potentially sensient and that very quickly moves us into the description of what is human and what is not and where we have historically systems of accelerated dehumanization for example that resulted in the Holocaust the when we uh move to models where we need to be thinking about sensient beings of which humans are a sensient being and um everything that we talked about earlier in this podcast with respect to AI sovereignty if we're talking about an AGI that is sensient as opposed to an AGI that can perform produce deliver things that that are outputs but is not a sensient entity a non sensient AGI can be self- Sovereign and will be and the idea of a general
purpose AI The General capabilities are already very rapidly expanding but if we talk about something that's sensient if that became self- Sovereign that would be a slave and we do not believe in slavery in our civilization any longer and that does bring us back it is a segue back into you asked for some you know what's going on practically right now how are we building toward this Vision if our overall principle is that personal data has become inter ated with personhood it just as every major technological advancement in the history of humanity has become part of our personhood our chief scientific officer that uh Dr Kahana knows well Dr sea Manion describes these devices as our extended neocortex and uh every breath we take even it isn't only my actions and deeds and words and choices it's the functioning of my biology itself all of which is increasingly part of the data that is exponentially increasing in every move I make every purchase I make every place I go Etc so one of the most
urgent things that must occur as a precedent condition to being able to implement living systems of ethical risk management with inclusive governance transparency and accountability Etc is that you have to implement the legal constr s that can begin to enforce and clarify all of the distinctions between these new relationships between individuals families and communities and the state corporations and getting that Baseline set of new institutions that's really what it is these are new institutions that are not reflected they you can see them in the negative space of society today where are the pieces so we are bringing about a set of new institutions that we purport can begin to go down this path and uh that involves also a whole lot of Technology lots and lots of technology that that can facilitate different approaches to machine learning to the uses of AI at runtime to especially systems of validation control and guard rails regarding the use of AI all moving quickly toward this new system boundary around the individual that can be the nucleus of
creating of creating that construct so this is you can't have new capital flows and New Economic distribution without establishing these constructs that can then receive and make and ride on new payment rails for so so it's first things first but fast as we can and uh there's lots of detail there that I hope that gives you a sense of it no it definitely does and I'm excited to see all the announcement that comes out on this line of work it's going to be incredible bit piece of work I think um you know we obviously this topic has been ethical AI but you are both at the center of web 3 in my mind and have many projects in the space I'd love to hear from both of you what are you most excited for the future of web 3 generally and longterm in the space well I I can start by say I'm excited by the project that I'm doing right now I uh together with a couple of colleagues we purchased a private island
in the South Pacific in the country of vanatu the reason we chose Vanuatu is because not only is it a paradisiac place but also it's been on the top even though it's a lowincome country it has been on the top of the happiest countries on the planet for the last 25 years so I had to go down and see okay why are you so happy making less than $4 a day and in that space actually we are applying as a social experiment all these Technologies these technologies that are driven by llms by L large action models that are principle based we're not a rule-based community we are schema based saying you know the only thing that needs to happen is that you bring more energy into the island than you take from the island and so we measure everything everywhere from water consumption to steps to how loud you speak and uh uh really try to create an environment where deep thinkers and doers like Heather and Sean and the whole team there can apply
those I think that what excites me is that and I'm kind of paraphrasing on Atul goand day this is the first time in history where we're doing wrong things because of ineptitude it's not because we don't know stuff that we're not doing yet we know stuff and if we don't know we can ask and that machine will know we don't we know because of we we don't know because of choice because we decide to be inept because we decide that it's convenient for us to be ignorant and that really brings us you know to the question of knowledge and what is knowledge and I want to say that it's exciting because for the first time you realize that there is no limits to what you can think about even if there are no words to express them it is incumbent upon us not to ignore it because when you start ignoring things when you start to create a reality that doesn't exist you introduce ignorance and you become ignorant and there's no reason now with
all the technology that we have and I always say this and Heather heard me say this dozens of times it's not the artificial intelligence that I'm worried about it's the natural stupidity love that line love that years Alex that project sounds incredible that is really impactful stuff I think um he that we are coming to the final moments of the show if you could just give us a really brief what are you most excited for in the web 3 space more generally that would be great sure I'll say that my my activity in web 3 has been about this intersectionality between web 3 and AI from day one um so so it has always been about that and I'm very excited because web 3 is the core mechanic through which we can create the institutions the guard rails the controls the safety the orchestration of we're talking about technology for computational trust technology for massively multi-party coordination technology for entirely new methods of systems of compensation using value exchange across any way that individuals wish
to Define that value there's the implications for freedom and liberty and civil liberties that were envisioned you know Alex mentions the Magna Carta um I I am an American Patriot and I also am very passionate about the ideas of Western liberalism across the board and the idea that if you look at a wide enough Arc we are coming full circle into the intentions of of extraordinary thinkers that shaped our age from many generations ago and that we are right now at the fulcrum of the realization of that kind of Liberty and agency and systems of governance and government that can enforce and protect and defend those systems that moves me profoundly and has me leaping out of bed every morning and web 3 is it's the heart and the foundation of all of that so so we could go Point by Point through every technology and tactic under the web three umbrella but I think our audience understands what that means and you all do so that's what I'm thinking about yeah well I think
that's a really powerful way to to state it and it's not often you hear someone with that much passion I think for getting up and and going to work every day and I think the whole idea around Liberty civil liberty personal Liberty self- sovereignty that is something that can you know get people out of B I think it's one of the Reon reasons why the web 3 movement has taken off so much so we're at the end of the show now and as you know as friends of the podcast we always end the show with two standard questions now I'm going to make an executive decision here I think we don't need to ask you one sentence what web 3 is to you I think you spent an hour telling us exactly what you think and if you want to hear a one- sentence version you can go listen to the both Heather and Alex's previous episodes but I am going to ask you the second question Okay so we always ask as you know
if you could choose anybody on the planet today from throughout history could even be someone fictional maybe someone from your June poster Heather in the background if you could find anyone to sit down and discuss web 3 with who would it be and why and I want to hear an answer that's different to last time so Heather we'll start with you and go to Alex Heather last time you said mahat mandhi who do you choose this time Frank Herbert Frank Herbert's a science fiction author and you may notice behind me you see that's a sand worm floating ominously over my head and this is a a a movie poster from a version of Dune uh the sci-fi classic that was never produced I am a huge fan of Frank Herbert and the Dune Universe and everything about it and uh that that world takes place notionally something like 10 20,000 years in the future of today and it is against a backdrop where 10,000 years ago icient AI had decided to attack humanity and humanity
went through a massive Thousand Years struggle to overcome and then had to create a new civilization and this new civilization is now 10,000 years old nominally that's the backdrop Frank Herbert was is an extraordinarily preent individual and I want to talk about how it is that we can realize extraordinary benefit for the future of humanity how we can prevent endless human suffering and how we can do it in a way that doesn't require revolutions and doesn't move at the pace of evolution but instead can be brisk and um directed by Heartfelt and Visionary leaders all over the world yeah I think that's a really awesome answer and obviously in June there's all these themes of sovereignty and you know States and things and how individuals work alongside side each other you know for the greater good so I think that fits perfectly with your your whole thesis on web 3 Alex how about you last time you gave us Emanuel lavenus who was a French philosopher so who would you who would you go for
this time at the dinner party so I think someone a little bit older than him his name is Edmund huser all this time he is a German philosopher the reason I chose him in the 19th century the reason I chose him was because he was a mathematician and all this you know thing about Ai and algorithms and math and all this so his passion was to explain the world through math and in that 50 year journey of trying to do that he realized that actually it's philosophy or the philosophy of math that really is explaining so I would like to ask him so what does he think about Ai and the philosophy of AI that is behind that is forcing us to think about personhood Humanity inter person and Inter sub experiences and thinking about the aboutness of ourselves and about world the wonderful world that is around us and I want to say that the thing that he came up with that I think is so uh unique is that he said unique is
not special because when I say that I'm special I mean that then you're not so you're unique Jack Alec Heather I'm each and every one of us is unique in other words unique is common what makes us human is that we're unique and that uniqueness will not only can but will be captured by these technologies that we coin them together as web 3 that it just solidifies this fidal world around us and reminds us that we're human this is what it's all about that's a great answer and it's another one to add to the reading list it seems every time Alex comes on I have someone else that I need to read up on so thank you for that this is the second time you've both been on I'm sure and I hope it won't be the last but thank you both for joining us and thank you to those listening wherever you may be and join us next time as we untangle a little more of web thanks for tuning in to another episode
of untangling web 3 don't forget to send us your thoughts questions and comments on social media and be sure to follow us on your favorite podcast provider to catch the next episode see you next time to untangle a little bit more of web 3 the views we express here are our own and do not reflect the views of our
Topics: AI & Robotics