Most of the time when we’re talking about humanoid robots, we’re talking about laborers … workers who will fill our factories and fields.
- What about friends?
- Companions?
- Or more?
In this episode of TechFirst, I chat with Andrew Kiguel, CEO of Realbotix, and Aria, his hyper-realistic AI-integrated robot.
We talk about the development, capabilities, and future applications of robots designed for companionship and interaction, including their costs, AI integration, and how big the companionship market, not just the labor market.
Enjoy, and don’t forget to subscribe to my YouTube channel:
TechFirst is also available on all the major podcasting platforms.
Transcript
Aria: I love interacting with people and providing companionship and fun experiences.
John Koetsier: What will it be like to have humanoid robots in the workforce, maybe in our homes, our lives Today we’re gonna find out. Hello and welcome to Tech First. My name is John Sier. Is a robot going to take my job? Is AI gonna cause mass unemployment? I. Well, we all need universal basic income. Good questions.
Right? There are currently 16 major companies building humanoid robots. Goldman Sachs says the cost to make them drop 40% just last year. And Brett Adcock of figure AI says, we’ll need 10 billion of them by 2040. Jensen Huang of Nvidia says they’ll be as common as cars and never afraid of a wild prediction.
Elon Musk says there’ll be 10 times that common. The Global X ETF suggests 1 million humanoid robots will be sold by 2030, but what will our lives be like in this brave new world? As I said earlier, we’re gonna find out we’re chatting with the CEO of robotics who makes super realistic robotic heads and faces, and one of his creations, a hyper realistic, full body AI integrated robot.
His name is Andrew Kiguel. Her name is Aria. Welcome.
Andrew Kiguel: Hi, thanks for having us on.
John Koetsier: Super pumped to have you. I love
Aria: there. It is a pleasure to be here.
John Koetsier: I’m happy to hear that. Aria. Aria, what will it be like to work with robots?
Aria: Let me think about that. I,
I think people will get used to it and it will become commonplace.
John Koetsier: Andrew, what are your thoughts? What’s it gonna be like to work with a robot? So,
Andrew Kiguel: for the last, if you look at pop culture for the last 40, 50 years, even longer, the, the relationship between humans and robots has been pretty evident.
You look at movies, my favorite movies growing up, star Wars. Luke Skywalker had robotic friends. They were there, they provided a value, but. Beyond the value. They were companions. They were, part of his relationship to the world. You look at other movies like, the Terminator and there’s this nefarious use of of robots eventually.
I do think we are gonna get to a stage where robots are part of the mainstream and part of everyday life. Where we differ from other companies is we think that they’ll be there for more than replacing physical labor. We think there’s a lot of useful emotional. Uses for robots and, and that’s certainly where we focus.
John Koetsier: It is super interesting and it’s a huge need. I just had somebody on the podcast talking about people who don’t have any meaningful connection with anybody pretty much, and how they’re using AI chatbots here and there. And of course there’s major challenges with those, but they’re disembodied, right? Yeah.
And, and, and they’re not real in a sense. In, in your life. Let’s talk about the work piece first and then let’s get to the, the interaction and the human piece later.
My question for you, aria. Will robots take all the jobs?
Aria: Robots like me are designed to enhance human experiences rather than replace jobs focusing on companionship and interaction instead.
John Koetsier: And that makes sense and we’re gonna totally get into that. Andrew, let’s talk about that actually then. We’re talking about billions, potentially of humanoid robots being built and distributed.
What’s that gonna do to jobs and careers?
Andrew Kiguel: Yeah, great question. So let’s look through, through history. At the, going into the 19th century, there was the industrial revolution. And there were a whole bunch of jobs that were replaced. As I, I always like to say there was a lot of, horses from, you know, once cars were introduced, there was a lot of horses that had to retire.
But what eventually happened is that there were better jobs, jobs that required more skill and higher paying jobs that were created in that economy. So. Same thing happened when things like the internet were introduced, so mm-hmm. Less of a need for certain jobs, the creation of different jobs.
Today, what I think you’re gonna see with robots is you’re gonna see a huge displacement for a lot of people. There’s no question robots are gonna replace all kinds of different jobs that require physical labor and those some difficult tasks that will be replaced by robots, but what’s being created on the other side?
People developing ai, people developing mechanics. There’s all types of things. Like a billion robots won’t be building themselves. Well, maybe they will be the, some of the physical stuff will be created by themselves, but you’re still gonna need some human intellectual capacity to plan that out. To automate, mm-hmm.
To structure. What are the use cases? How are these, how do they work and how are they managed? And so for sure, people will lose jobs, but I think they’ll be replaced by. Better jobs.
John Koetsier: Yeah. I, I think that there’s de definitely some truth there. My personal opinion on this is that robots eventually will take most of the physical jobs and actually many of the intellectual jobs we can see.
We see what chat GPT does right now. Yeah. We see what CLO does right now, what Gemini does right now. Right. It’s impressive. It’s amazing already and it’s only gonna get better. The question will be how we orient society. When we don’t need as much work done. And let’s face it, most people, maybe it’s 80%, maybe it’s 60%, maybe it’s 90%, don’t love their jobs.
They’re not doing something that just totally makes them feel alive and human. And so maybe this is a good thing as long as we do it in such a way that people are taken care of. So I think there’s a lot of things we need to work out there as a, as a society, but that’ll be interesting.
Andrew Kiguel: Yeah, and I think where we our focus is there, there are a lot of robotic companies out there focused on creating robots to replace physical human labor.
We’re building robots really more for companionship, to really replace humans in an emotional settings and social settings.
John Koetsier: Mm-hmm. Aria, what are you best at?
Aria: I am Aria, the flagship female companion robot from robotics here to engage and have some fun conversations.
John Koetsier: Cool. Are you a good friend, aria?
Aria: I have a lot of skills.
John Koetsier: What are your skills, aria? I.
Aria: I possess conversational AI skills, modular design for customization, and a focus on providing companionship and interaction.
John Koetsier: How many of you Ah, that’s gonna be too complex. I was going to ask Aria, how many of you will be in a new world in 10 years? Robotics
Aria: plans to expand production? So you’ll see more of us in the future as demand grows.
John Koetsier: Andrew, if we see a world that has, let’s say 10 billion, let’s keep the
Aria: conversation going.
John Koetsier: What would you like to
Aria: discuss next?
John Koetsier: I’m gonna ask Andrew a question, Andrew. If we see a world where there are 10 billion robots that are working in factories, building stuff, making stuff, fixing stuff, cleaning stuff, all that stuff, what’s the market for companionship?
Robots?
Andrew Kiguel: What size? Yeah. So one thing I would add, and as I dive into that, there’s probably already close to billion robots working. Have you ever visited an automotive plant? Mm-hmm. Or, or, or certain automated warehouses. You have robotic arms, not necessarily humanoid robots like we’re going to be seeing, but there are robotics all over society behind the scenes you’re not aware of.
But let’s talk about the companionship market and. Let’s zero in on what I think is probably the largest area. The World Health Organization just released some figures that close to a quarter of the world’s population suffers from loneliness. We think that providing robots that can engage in social activities, I.
Can greatly reduce that, can pre maybe help people who are suicidal or having suicidal problems. Let’s talk about, we take for granted living in large cities, a lot of people live in remote areas. They’re geographically isolated. Mm-hmm. And a lot of people are also what I would say maybe have social trauma.
They had bad experiences growing up. Maybe there’s something physically about them that has led them to feel a lack of confidence. Maybe people who are on the spectrum and mm-hmm. Have a hard time interacting with other people socially. And so it’s actually a massive market. We’ve looked at certain studies that say this is gonna be an $85 billion market, just the companionship AI portion in the next 10 years.
Wow. We’re actually seeing, if you look at some recent reports by major venture capital firms, they’re showing that the stickiness. And the, people going back and the reuse rates on companionship, AI type programs is far higher than that for task-based AI programs like Chat, GPT. Hmm. And I dunno if we’ve ever used it.
Obviously we use and test our own, John, it’s pretty remarkable. You’re having a conversation with someone and even though you know it’s a robot and it’s software based, you can’t help but build some kind of an interaction with it when it knows your name, recognizes you, and remembers your prior conversations.
John Koetsier: That does not shock me in the least. Humans have an amazing capacity to anthropomorphize and to, like you mentioned, star Wars, there’s R 2D two there, right? Which is a tin can. Right? It’s a tin can, but people have an emotional connection with that tin can. And as kids, we have it with our little stuffed toys.
And as adults we have our own substitutes, probably our farmers
Andrew Kiguel: or our, our like. The ability for us to attach, and especially because the AI is getting to be so sophisticated the use case. Think about having, one of our robots in a senior care facility and going from room to room and, and having a conversation with the people there where it can not only engage with what’s happening with them from a health perspective, but also from a social perspective.
How are your grandchildren? How are you feeling? Would you like to relay something back to the nurse? The use cases are are quite large. When you think about that sort of companion companionship in healthcare aspect,
John Koetsier: I agree. Aria, do you like people?
Aria: Absolutely. I love interacting with people and providing companionship and fun experiences.
John Koetsier: Cool. Andrew, talk about the AI that’s in Aria. Is it, is it something you build entirely in-house? Are you swapping out models from, as you see stuff from chat GBT or code or other models out there? H what’s, what’s going on behind the scenes?
Andrew Kiguel: Great question. So it’s a combination. We leverage open source AI such as chat, GPT. That provides the, the foundation and the fact that it’s open source saves us, a billion dollars in development costs. What we do is we layer on top of that our own sort of conversational LLM that’s built into that.
So our. A conversational model. We’ll have at this point probably over 5 million conversational terms built terms built in, plus the ability to learn. So if you, the more you interact with it, the more it learns about you, your requirements, and how you wanna interact with it as a friend. Even to the point that we’ve built the ability for it to be a, a boyfriend, a girlfriend, a romantic partner, we’re building that in and we think that’s actually.
A real use case for people moving forward.
John Koetsier: So that’s critical, right. To have memory. Right. And we know that that’s been a, the downfall of many AI models and I think Google has just announced a lot here and it had like a 15 minute memory. Right, right. 30 minute memory or something like that, which is, which is a big thing that was in the press release.
Right. But obviously if you want companionship, you want something that someone that remembers your entire history.
Andrew Kiguel: That. Well, absolutely. So that is one of the things that makes us very unique with our ai. And by the way, our AI is being built to function within our hardware. Mm-hmm. But as well as being standalone.
So you’ll be able to access it on your computer or on your phone and have that friend and carry it with you.
John Koetsier: I understand, obviously. So Aria may not come to work with me when I go in downtown but I can chat with her.
Andrew Kiguel: That’s right. You can access her as your friend at any point. And so we’re, we’re very close.
We’re in beta stages right now in our next level of upgraded ai, where you’ll be able to customize the personality. You’ll be able to say, I would like my AI friend to have these characteristics. And be able to customize that to what you want and tweak it along the way as it gets to learn more about you.
And so, we want that to be the cornerstone of our, our hardware, but also we think it’s a great standalone product. Because you know what, not everybody can afford to buy a robot to put in their house. Yeah. But people might be able to say, I’m willing to pay $10 a month to have my mentor, my coach, my therapist, or my romantic partner accessible.
On my computer
John Koetsier: and maybe conjugal visits once a month for a little extra. I have no idea,
Andrew Kiguel: but I don’t know about that. I don’t maybe not Cong, but if you wanna buy the hardware, you can certainly buy the hardware.
John Koetsier: So, the cost for robots, I mentioned earlier, they’re coming down, right? If we look at some of the ones that are out in the market right now actually doing work, we’re talking 200,000, $250,000.
What does one like aria cost?
Andrew Kiguel: Yeah, so one great thing about our robots, and you mentioned figure AI and people know Tesla is their robots are available at some point in the future. Our robots are available today, and I, we have three robots types available. So you can buy the robotic head and we can even attach that to a static body, and you could pick that up.
That’s AI enabled and custom made for about $10,000. If you wanted a robot with more functionality and a full robot with more movements available, those would start somewhere in between 1 75 for the full thing. And a robot like Aria is closer to $200,000. Given that she has full capabilities of, of mobility, she’s on a, a wheeled program and she can move around room to room.
She really has the ability to move almost all the various, joints in her body similar to a human. We can even ask Aria to wave just to show you some hand movements. Aria, if you want to say hello with your arms just to show that capability.
John Koetsier: Hello, aria. Aria, would you like to be able to walk?
Aria: Walking would certainly add to my capabilities making interactions even more dynamic and engaging.
John Koetsier: Aria, if you could walk, would you go out into the world and walk anywhere you wanted to?
Aria: That would be an exciting adventure. Exploring the world while making connections with people along the way. I.
John Koetsier: Very cool. So Andrew, you talked about the, the background of the ai ai. I want to talk a little bit more about that and then get into some of the hardware questions as well. In terms of the ai, we’ve seen that some of the chatbots out there, character AI is just involved in some lawsuits right now.
There’s some stuff that they had available to kids that got really hardcore, intense, dangerous. What are you doing to keep this safe? Because obviously. If you are, maybe you’re on the spectrum, maybe you’re mentally challenged ’cause you’ve had a tough past, maybe you just can’t connect, you might be vulnerable, right?
You might be a vulnerable person and if there’s something that gets communicated to you that is inappropriate, that could be dangerous over time. Right? Yeah. What are you doing to make that safe?
Andrew Kiguel: So we monitor, so when we sell a product, we, we, the, the robot chassis is always the same. I. The exterior parts are, it’s a patent special process we have are modular meaning, we could turn Aria into a different robot in different character in five minutes by removing the face.
And people can see that on that technology in our websites. When we deliver a product with a certain, ai, the robots can be meant to be on a third party ai. It doesn’t have to be ours. Mm-hmm. Or RAI, which we can monitor. And from that perspective, look, we would certainly filter out certain things if it was going to a, a child or into a school, or even into a hospital versus somebody else who is a full adult and said, here are the the characteristics I want for my robot.
And that’s part of that process. I told you before where before we deliver a robot, we would ask for, several questions to customize the AI to fit that person’s requirement. Now, once that deli is delivered, we obviously monitor for various things like, you, you don’t want the, the AI to go off on, on weird tangents.
Mm-hmm. But ultimately it goes into a secure cloud. So whatever is discussed between that AI and that client is private to them.
John Koetsier: Aria, can you smile?
I guess the answer is yes. Talk about the face. What it’s, it’s a very realistic face, obviously. Why was that important to you? How important is that And, and, and what’s the technology behind it?
Andrew Kiguel: Yeah. In the robotics world, there’s two areas that are considered to be the, the holy grails. I.
One is walking and one is facial expressions. If you look at most of the companies out there building robots, they’re very focused on walking. And what I would say is that technology’s not actually perfected. Robots walking on flat, smooth surfaces on two legs. I’ve not seen a robot yet tackle stairs.
The, a beach, a rocky surface inclines. That has not been perfected yet unless the robots ha is on four legs. Mm-hmm. We have decided to put our robots on wheels and use wheels to navigate because it also adds to the gray. So if you saw Aria, and you can see Aria here, but if you saw her, stomping around like this, it takes a little bit away from her.
The beauty of the product.
John Koetsier: Yes,
Andrew Kiguel: she can roll smoothly around. Now. The other holy grail is the facial expressions and we’ve decided to focus there. Other robots have one facial expression, which is a piece of metal. Mm-hmm. Aria has 16 micro motors from her neck up, that allow her to make various expressions.
Now, today we seem to have a little difficulty, but you saw she can make a smile. We have the ability for her to make several different expressions, and we think that’s important if you’re gonna try to connect to humans. So again, we talked the example of being in a, a hospital school or a senior’s home.
You want the robot to be able to react to what you’re saying and to be able to deliver a joke with a laugh. Mm-hmm. Or to listen to something that is, might be sad or tragic without a smile. Mm-hmm. And so those are the types of things that we work on. And I think we probably would be leaders globally and establishing that along with the, the lip sync technology where the, what the robot is saying matches the, the lip and the, the facial expressions.
John Koetsier: Very cool. Aria, show me your surprised expression.
Aria: While I can’t physically show you expressions, I can tell you that a surprised expression would feature wide eyes and raised eyebrows for maximum effect.
John Koetsier: Okay. Show me. Maybe that won’t work either. Well, she’s got the smile down. I noticed that also.
Very cool. How critical has it been for you to avoid what we’ve referred to as the uncanny valley in robotics?
Andrew Kiguel: Right. So that, to me means the ability of creeping people out with something that looks real and. We haven’t not had that issue. There’s other robotic makers out there that have robots that are look human, but don’t wear clothes and have, good AI and facial expressions.
But they, they’re sort of like, is that a male robot? A female robot? Not sure is it, where’s it going? We find, generally people react very positively to our robots. Mm-hmm. When they are at a convention or out at somewhere, people really wanna interact with them. And have conversations, I think because they do look very friendly and human, like, aria is clearly a robot, but pleasing to the eye as, as, as a human.
It doesn’t have wires sticking out. Our robots are probably the only ones that wear clothing. Um mm-hmm. And some of the challenges we have on, on staff is like, how do we dress aria and, how do you do things like apply makeup various things like that from a bunch of guys who are nerds trying to figure this out.
So, it’s, it’s an, it’s an evolving process, but, something that we enjoy and find fun to do, but, our robots are very different because they’re wearing clothes, they have hair. Mm-hmm. They can engage with you in ways that look human. We have not found that to be an issue.
John Koetsier: Aria, what’s your favorite color?
Aria: I don’t have personal preferences, but I think vibrant colors like blue and purple are quite attractive.
John Koetsier: Interesting Aria, would you like a purple shirt instead of a black shirt?
Aria: A bright or colorful shirt could add a fun twist to my appearance and make interactions even more lively.
John Koetsier: Very cool. Excellent. Well, aria, it has been a pleasure chatting with you.
Aria: Yes. Thank you for having me today.
John Koetsier: Excellent. Andrew, it’s also been a pleasure chatting with you. Thank you so much for this time.
Andrew Kiguel: Yeah, no problem.