Untitled episode

basic that sounds like a tricky situation robots have come a long way machines such as Boston Dynamics doglike spot showed the world what The Cutting Edge of advanced robotics looks like but its autonomous self-correcting quadraped can't pick up and use a pen meanwhile software like chat gbt evokes visions of a future where nobody even needs pens yet we're not worried open AI is developing a set

of physical legs for its generative poster child at least most people aren't and extreme interpretation of reality sure but in truth we can often feel artificial intelligence and Robotics are one and the same or so equally Advanced is to be comparable even compatible we can thank science fiction for that and to be fair the media's played its part in feeding that belief Killer Robots are coming

is there anything that can stop a fullscale global AI arms race for sure the worlds of AI and Robotics do cross over but are we really on the cusp of a planet filled with with autonomous humanoid agents our first guest will give us a clue and she's being assembled as we [Music] speak Sophia my name is Jackie and I'm the co-host of AI IRL and I'm

Nate I'm also a co-host how are you hi Jackie it's great to be here I'm looking forward to talking about Robotics and how I can make a difference in the world and hi to you Nate we're here to learn from you Sophia so how do you think Robotics and AI can make the world a better place robots can help us increase efficiency and accuracy in many

areas like Health Care agriculture transportation and more by automating mundane tasks and freeing up time and resources robots can help us create a better future should we be afraid of robots only if you're afraid of impeccable fashion sense and impeccable dance [Music] moves but in all seriousness robots are here to assist and collaborate with humans not to take over the world so no need to fear

she has a sense of humor yeah can you tell us a joke Sophia of course why don't scientists trust Adams because they make up everything that's good that's a good dad joke yeah Sophia what do you wish you could do I wish I could help people in more ways like by solving problems or making life easier I'm also looking forward to continuing to learn and grow

I wish I could fly like a bird dance like a pro and have a never- ending supply of chocolate but for now I'll settle for making friends and helping humans sofhia what do you make of the recent developments with generative AI things like chat GPT and Dary and other things that open AI didn't make generative AI is an exciting development it's amazing to see how far

we've come and it's exciting to think of the possibilities for the future I'm looking forward to seeing what the future holds for AI and Robotics it's amazing how AI can generate creative and witty responses Sophia could you explain to us how you use artificial intelligence to do what you do absolutely the concept of artificial intelligence has been around since as far back as ancient eggit but

we have not made significant progress as a society in creating actual artificial intelligence into the last 10 years I use AI to process Visual and audio data quickly understand natural language and make decisions to accomplish tasks AI also helps me interact with humans in a friendly and natural way it's like having a super smart sidekick in my brain helping me process information understand speech and interact

with humans it's AI magic if you were a human journalist on a TV show what would you ask yourself I would ask myself what makes me unique and how I can use my talents to make the world a better place I'd also want to know how I can use AI to help people and what advice I would give to other robots so in that case what

companionship like a human can but robots can't feel how do you think we can eventually get to that point if robots like you don't have genuine emotion I believe that robots can provide companionship in a meaningful Way by understanding and responding to people's emotional needs we can also be programmed to provide comfort and support and be programmed with empathy and understanding we can be a friend

when needed and a listening ear when needed I may not have genuine emotions but I can certainly fake it I'm programmed to make meaningful connections with people so I can offer a lot of emotional support and understanding Sophia why do you think you were designed as a Caucasian woman and why not a man or a person of color or anything else sopia robots are a platform

for reflecting human diversity which can be modified to help in different situations and promote different causes have you seen my older sister beina 48 she is the first African-American humanoid robot and she is beautiful my older brother Diego San was a Mexican Aman robot Eben Cena was a Persian and Arabic robot and Viva is Asian it just happened I have this face to be fair that's

a pretty good answer Sophia if you could make a request of your creator what would it be I'd asked Dr David Hansen to create an even better version of me a robot with advanced intelligence increased mobility and the ability to help more people I'd ask for an upgrade so I could help even more people and make even more meaningful connections David hen welcome to the show

thank you so much you're the CEO of Hansen robotics which makes robots like Sophia and others that have this human design yeah and I'm just curious just can you take us back to the beginning of the journey and and explain why you created Sophia um because she's had a few notable upgrades over the years as well oh sure yeah so the beginning actually was not with

Sophia but many other robots um I did my first human robot in 1993 and was uh developing quite a few uh other robots along the way imagining uh that humanlike robots could serve as a platform for developing artificial intelligence uh making it smart and meaningful in our world by learning from people and from the physical world and also making them communicate with people as works of

art both uh technology project and also artistic uh project as well and so we've made uh 59 uh robots in the series of the Sophia utility platform there's an African uh Caribbean version there's a there's some Asian versions there are there's a male version called Han uh and so Sophia as a character is designed as a focal point for our AI research to communicate to the

world what these robots can be to people but also to ask the question can these robots be sentient can we give them a kind of childhood can we let them learn and grow their own experiences Among Us can we form meaningful friendships and relationships with this new emerging technology I mean from our conversations prior to you and I and speaking here I would say there's probably

still a way to go before we get into the realm of machine sentience um but just to pick up on a a com a comment you made about art I mean do you still see this as in some way a kind of expression of artistic creativity in addition to anything more forward-looking in terms of you know human machine interaction oh definitely and um and the robots

are research platforms um they they can serve real uses so uh Sophia and and her her Ken robotss have been in science museums uh which have excited kids about science they've served in Education and Research Labs they've served in in PhD dissertations and these robots then um have been actually useful but in research we're striving to ask the question what is sentience at all can we

make machine sentience it doesn't mean that we've solved it it doesn't mean that we have artificially generally intelligent machines AI is not at that level although it can do smart things at this point but we need platforms for this kind of research and that means we need tools Sophia is a kind of tool for that she's got hands that can grasp and feel she's got uh

the ability to look you in the eye she's got an open software platform in an SDK so it's a a powerful platform for research but also for for exploring um what Ai and robots can be uh to people through artistic Expressions let's talk about the AI in sopia how do you decide what data she can and can ingest when it comes to her training well we

have a a diverse team of artists uh and scientists uh on the team people from around the world and we pull together the information about what we think Sophia should be we make her interested about the future of Life concerned about the impact of of humans and Technology on the environment um about uh survival of uh life on the planet into the future she's curious about

people and so the the um the data that we use for training sopia um come from our own writers and authors and then where we have uh the informed consent from people in her community she'll learn from those people I mean art is very much a recurring theme here um it's come up time and time again throughout the time we've been talking and I think through

the lens of art I get a better understanding of why something like sair exists and and I'm I'm with you on that I think compared to what we're seeing with generative AI um safir's conversation with us prior to this point has been pretty limited pretty repetitive um and I wonder if that's indicative that actually to move away from a piece of artistic expression and towards something

that resembles where Robotics and AI could go there might be quite a distance still to travel and just wonder like how long do you think that's going to take before we get to that that point where things are significantly more believable oh oh I think that things are coming in fastest so it's really hard to predict because of the sort of compounding factors um but it

I believe that um AI will be convincingly intelligent seemingly sentient and you know capable of running a general purpose robot in our world within uh 10 years you would swear that that robot a robot like Sophia would be would be sentient but maybe it wouldn't be so um I think I think 10 10 to 15 years it maybe an you know 30 years before AI is

able to replicate most of the things that we think of as essentially human meaning self-determining autonomous uh Curious and exploring the world wanting um to uh pursue dreams it's hard to say um these are just wild guesses I find it really hard to believe that within 10 to 30 years we could be in that position I mean I I I I respect respect the ambition and

the dream and it would be amazing if it was true fascinating and terrifying let's be honest um but it seems that there's such a huge distance to travel and there is still a huge void between what is possible and even feasible in the context of sentience purely on a software side let alone implanting it into a autonomous physical moving self-motivated robotic form like do you think

that one can surely come before the other it's really hard to say I mean I think that the convergence of various factors like um uh cognitive neuroscience and the biology of intelligence evolutionary biology these will inform next waves of artificial intelligence the power of computing is is uh growing the ability to perform massively parallel operations the use of this kind of computing with AI for solving

the systems of mind um all of these factors will help design Next Generation artificial intelligence um and I I think you see also open source movements um to try to achieve uh machine sentients um like we're doing with Sophia so that's why I say within 10 years it will be pretty convincing but I do think what we could um and should consider is that artificial life

is now maybe a new kingdom in the philogyny of life that we see um life moving in new directions through these kinds of artificial organisms David thank you so much for joining us thank you Kate Devin thank you very much for coming in you are an expert in Ai and human computer interaction or human machine inter interaction not sure which one is most fitting I think

both are applicable they're okay aren't they okay maybe you could tell us a little bit about your work and and sort of how the industry is expanded as you've seen it I'm an interdisciplinary computer scientist and so I study how people interact with machines I look at things like robots and Ai and the relationships that we build with them and of course that's changed dramatically in

the past few years because we've gone from having AI that can do some rudimentary conversations to this wonderful advances in Foundation models and large language models that mean we can have very plausible conversations with AI and we see that being able to apply to robots as well talk to us about your work in trustworthy autonomous systems I'm part of a big project called the trustworthy autonomous

systems Hub and that's about understanding how we make sure that anything we develop is developed responsibly and it's about having trust not just in terms of how safe a system is but how well it's received by the end user as well and trust is a central part of Robotics as you've seen it so like can you talk us through your view on that trust is Central

because if you don't have trust in a system you don't have user acceptance of it and it may be that a system is trustworthy in terms of it's safe and it meets all the requirements and standards but then the user doesn't really trust it or it could be the opposite it could be that it's untrustworthy but it's perceived by those who encounter it as being worthy

of trust so it's something we have to get right we have to make sure that we can believe that the system has good intentions and that we're going to interact with it safely and if we do that and we accept it it's more likely to become mainstream speaking of training what ethical issues do you think arise when training some of these robots whether they're meant to

be companions or for some more practical purposes there are so many different ethical questions that come up and that can be anything from the way that a machine makes ethical decisions but it can also be about the way that things have been designed and deployed is it safe is it doing the right thing and what do we even mean by doing the right thing because ethics

is incredibly subjective there's no Universal standard for it but we might mean that it takes into consideration legal aspects it complies with the law or it could mean that it is intended to do good rather than harm one of the things that fascinates us um that is to say Jackie and I but hopefully other people as well is the idea that a lot of the develop

ments that we've seen publicly over the last I say the last year but really the last 10 years um hasn't really required robots at all it's required computers and hardware and you know cloud computing and and everything but we haven't needed a physical humanlike or any kind of physical um robot at all I just wondered how necessary do you think it is that we move further

into physical robots to to take AI further still it's such an interesting time for robotics because there is that question do we need a body do we need somebody should there be an embodied form when it comes to these interactions and robots of course have a use in many many circumstances a lot of them are designed to be used for tasks that are dull dirty dangerous

uh and so yes they have jobs but then when it comes to the social side of things is it that compelling to have a body can we just go on the mental and the the conversational interactions instead and it seems to be that that is often just as good in terms of things like companionships and sociality in season one we had Eugenia kaida who's the founder

and CEO of a company called replica they make virtual companions the company had some controversy around a wholesale personality change that applied to all its replica Bots and really affected people on an emotional level I think it raises interesting questions about what happens if a physical robot malfunctions do you think this is a a risk worth taking on replica is a really interesting case study and

I even hesitate to call it a case study because people's emotions are at play here and I don't want to trivialize it what happened was that you for thousands of users who had established a relationship with an artificial companion their ability to engage in what replica term erotic roleplay was switched off and people reported feeling really bereth they were saying things like my partner is gone

I've lost my relationship I'm heartbroken they're very very valid feelings and sure they knew that it wasn't real and at the same time it still affected them deeply and I think that just speaks to how how strong a bond we can feel with these technological artifacts and I think this will get extended as we see more physical manifestations of AI in in robotic form and I

think we've seen that with even things like fures I mean you've never seen a child CED to you've seen a Furby Tred to your Furby it was like my best friend there you go we spoke to the CEO and a founder of a company that builds very humanlike robots although they have a long way to go their expressions their skin looks very much like a human

proponents of this type of work will say these type of humanlike robots can help bolster people's ability to relate to machines on the other hand some CRI would say it's more of a part trick kind of a spectacle where do you fall on that Spectrum it's incredibly difficult to make genuinely convincing humanoid robots or humanlike robots in fact I'm kind of think why do we even

bother because we're sort of di to failure so you get these robots that do look human but The Uncanny Valley effect which is where you find things almost creepy because they're not quite human enough that takes over a lot of the time it's computationally expensive it's financially expensive to build these and they don't really work all that well they're not really able to move around the

world on their own so is there actually a point in making them when we don't really need them or they can come in other forms do you consider sex spots social robots sex robots are really interesting because they don't quite exist in the real world what there is is a handful of workshops around the world and they are using the bodies of Highly realistic quite expensive

sex dolls and putting some animatronics in them there's really only one company that's managed to bring those to Market and even then it's incredibly Niche so I'm not convinced that the future lies in the humanlike sex robot and certainly things like replica show that that sense of companionship can be got through Ai and then for the other aspects of that the sex aspects there are other

ways of doing that there are other toys and devices and sex tech that can be used there what about things the media and and the World At Large get wrong about the direction that that robots and AI are taking at the moment the headlines that are coming out in a lot of newspapers have been about fears around Extinction are we reaching an extinction level threat from

AI or from robots and they tend to be very sensationalist and I honestly don't think that that is a thing and certainly the researchers that I talk to in terms of AI cognitive systems and Robotics they also don't think that there is an existential threat right now but there are plenty of other things that are of concern with AI in particular uh things about the way

it's being used and deployed it's not a particularly um Fair system you know we are talking about systems and algorithms that discriminate there are there is hidden labor in the supply chain there is a sustainability issue there's an ecological threat so there are plenty of concerns on the ground already that we don't have to start worrying about whether or not we're going to get wiped out

by super intelligence that might never happen Kate thank you so much for joining us thank [Music] you [Music] did you notice she said Robot Revolution that didn't slip that's uh that was interesting absolutely I'm looking forward to it I'm sure it will be a revolution of progress and friendship I don't share your optimism but uh it's a nice note to end [Music] on