Silvijs Seres: Hello, this is Lorn.tech. My name is Silvija Seres. And I am here today to talk about social robotics with Olli Ohls from, a company called Futurice. Welcome. 

Olli Ohls: thank you very much, Silvija. Great to be here. 

Silvija: Silvija. Seres tell us a little bit about who you are and then we'll dive into this really fun field you work in.

Olli: Sure. So I'm only working as a robotics lead at a digital design agency called Futurice in Helsinki and also in Europe, we operate. And so I'm working in the field of robotics, mostly focused in the social side of robotics or social robots.

Silvija: Can robots be social?

Olli: Sure. I mean, I think if you look at social robots, the social aspect of them is the most interesting aspect of them. And the concept of interactions is very, very like central to social robots versus.

Silvija: So you have to help us with some pictures here in our minds. I've been reading, for example, about Japan's enormous investment wealth into the area of welfare robotics. Basically, they have a huge demographics problem. Lots of old people and very few people to care about them or to take care of them. And then they give you back quite a lot of your investment as a company. If you find good ways to provide robots that will do, you know, health care or be a receptionist in a hotel or even a little robot seal. Why would one do that?

Olli: Well, I think you already mentioned a part of the kind of challenge that they're trying to solve is that they will have a lot of elderly people that need to be taken care of. And in Japan's rubber, this Asian strategy that they released a couple of years ago. Their state that the health care sector is one of the key areas where they want to create new innovations in robotics to actually solve. So one of these this challenge that they're facing. So one example is that, you know, how to keep elderly people in this care homes kind of active, for example, communicating more with each other, how to help with this like, you know, memory recall issues. So these are the kinds of things that the robotic scan can help with on social robots.

Silvija: I was very fascinating because I was thinking, you know, it's a furry little sealed robots and it somehow helps the people with dementia. 

Olli: Yeah. 

Silvija: Remember things just by sitting there making sweet noises and being, you know, cuddled with. 

Olli: Yeah.

Silvija: Does it react to voices? Does it? I mean, does it do anything other than a toy that you buy with toys or what is the social side of it?

Olli: Yeah, I think for this seal robot called Porro, it's more about like getting comfort, you know, it's like, you know, petting a dog or something you get you feel somehow empathy or something. This is the kind of thing. So the actual seals don't help you remember. But there are other applications that are being created at the moment in this elderly care homes where they actually were to remind you, for example, who are your family members by showing images and sharing their names, if you forget or like, for example, your items, like your keys or your phone. Robots can help you, you know, remapped like help you remember what these are and what you're supposed to use them for. And then also an interesting case was done currently in Finland where they found out that this kind of pepper robot did actually increase the amount of interaction between those elderly people. So it was not the interaction with the human and the robot that increased, but the amount of interaction that happened in the social environment among 

Silvija: somehow it triggers communications.

Olli: Yes. 

Silvija: You mentioned Pepper.

Olli: Yeah. 

Silvija: So that's another social robot. 

Olli: Yes. 

Silvija: It's been a very popular guest at conferences. 

Olli: Exactly. 

Silvija: What does Pepper do?

Olli: So pepper is a typical what we call a humanoid robot. So it has some human features. It has head, arms. It can dance around. It can speak. It can play songs for you if you will. So it's a very typical of what we call a humanoid robot. And Pepper is probably the one that you've seen in the media as well. Mostly this white robot kind of looking sympathetic and in some ways as well.

Silvija: But it's being oversold if you ask. 

Olli: Yeah. 

Silvija: Because people say, you know, it understands what you say and it will look happy or sad 

or but it actually has to be programmed quite carefully and it only reacts to a certain number of things. So tell us, I met you at an USSLER business forum. 

Olli: Yeah. 

Silvija: And you were showing the projection, I think, of Obama's face. 

Olli: Yeah. 

Silvija: On a face. And it could somehow smile to me when appropriate. 

Olli: Yeah. 

Silvija: How does one make robots understand humans or what do you do? What? What how do you implement the social robot?

Olli: Yeah. I think that's a very good question. And we can also think about is that actually understanding or is it just reacting? Is it a response? Is it a pre-programmed response to something like a human facial expression? And then the robot is thought to actually react in a way that we expect humans to react to. The case that you mentioned that we did with this, we call it Robama robot. Robama was just a demo of …

Silvija: Robama.  

Olli: Yeah, robot Obama is using this technology from this company called Ferhat, which is based in Stockholm. They have created what they call the world's most social robot because it it's not actually a mechanical robot in terms of the technology, but it's a projection. So at this makes possible this micro-expressions and much more expressiveness in the actual face of the robot. And this makes it possible to apply it in different kinds of areas. Then a more, you know, generic paper where you just have two eyes and a mouth, but they don't actually move.

Silvija: Mm-hmm. What about. So there are some where it's just a projection, a flat surface or less. And then there is this robot called Sofia. 

Olli: Yeah.

Silvija: Is that just please remind me. She has a face. Right? 

Olli: Yeah, 

Silvija: and she has some muscles.

Olli: Yes, that's true. So Sofia is what we would categorize as an android. So how that is different from a humanoid, is that an android? The creators of Sofia tried to make it as humanlike as possible, whereas in humanoid robotics we want to take some part of human interaction capabilities or appearance and apply it and into the product. But in Androids, we are actually aiming to create exactly human like behavior, appearance and voice and so on.

Silvija: So she can be touched then and it's sort of some material rubber that almost feels like skin.

Olli: I guess that's the end goal, that's what it should be by the definition of an android. I'm not sure how much in Sofia they have actually thought about the tactile feedback kind of that people get when they touch Sofia, because that would require then having some kind of warm, you know, human body temperature system.

Silvija: changes of color.

Olli: Yeah. Yeah, but yeah. 

Silvija: But she moves her face.

Olli: Yeah. 

Silvija: All of it. Or just the mouth or. How does that work?

Olli: I think in Sofia, which is, by the way, one of the most advanced robots or androids in the world. So she can do some facial expressions. And they tried to mimic human facial expressions as well as possible. I think they're not quite there yet. So there's some work to be done. But I have to say it's fairly sophisticated in terms of.

Silvija: So can I ask you, does it have to have a body? Does it have to be mechanical? I mean, you said already it can be a projection, but could it be just the voice? Is Alexa or, you know, Google home? Robot?

Olli: I think that's a very good question. And we're kind of the borderline and there are different definitions for it for a social robot. I think that a social robot should move in some way. But some people actually categorize this voice assistants that are in homes, that the voice comes out of some kind of box as a social robots as well. I think then we're only fairly limited in the interaction modality as to the actual speech and maybe some blinking lights, whereas we can go further by applying different modalities. And then only this one. 

Silvija: So I want to ask about something else. 

Olli: Yeah. 

Silvija: Why are we doing this? I mean, I am one of the first memories I have about reading about robotics was this description of the uncanny valley where you make things that are almost, you know, as human. And then we've detect small mistakes and it freaks us out. So it's almost in VR, it's almost better to deal with STICKMAN than with something that's, you know, trying to be human, but not quite is it because you want to manipulate people more. Or is it because they will react better to do things that understand them or what? Why do you care?

Olli: Social robots I think once we move from the two dimensional virtual avatar space into actual physical robot, we have risen the social engagement of people interacting with robots. So that's one key you know, difference with this kind of virtual avatars. And I guess that's one reason why these are being built. But also another aspect is the fact that robots can do other stuff in the physical realm. Many tasks that humans normally do in their everyday life like, you know, getting their coffee cup or whatever, it's like operating in the physical space. And I think now there's starting to be a kind of a movement where we're starting to see that the social interaction side of robots converging with the actual more industrial, you know, techie side of, you know, pick and place kind of robots or industrial robots in general where this these technologies will start fusing. And that's where we'll also be seeing more exciting applications of like robots that, for example, work in the elderly home that can pick up objects and, you know, tidy the home and help the elderly people while they're also conversing with them by speech. So this is very interesting time to be working with robotics.

Silvija: And I guess there is the whole world to explore when it comes to these human robot interactions. We don't know how to deal with these things because we haven't had them before.

Olli: Exactly. Yeah. And I think also related to what you mentioned earlier was the uncanny valley effect. So we don't really know what happens once we have, you know, interacting with robots for a longer period of time. Because there are some early research being done where the actual the eeriness that we feel towards these robots actually goes down in time. So that would indicate that we would get used to these kind of new agents working in this environments with us.

Silvija: I also remember you telling me a really interesting project that you worked with that had to do with autistic children and social robotics. Can you please explain that?

Olli: Sure. So we were approached by the hospital district in Finland, the West Coast, and they were looking for a robot that could teach sign language to autistic children in this rehabilitation center. And we found out that actually because we had built a humanoid robot life sized robots, and we found out that by doing some modifications on the on the design of the robot, we could make this happen. And this became a very interesting case, my colleague, Minja Axelson the design lead here. And we found out that actually it was quite an interesting experience for the youth and the children to be learning the sign language basics by the robot. So it was a small group, but the initial results were very promising.

Silvija: I'm just wondering, because, you know, there is a specific user group here. It was about autistic children and you. 

Olli: Yeah. 

Silvija: And maybe there is a different way of interaction that is superefficient for them. 

Olli: Sure. 

Silvija: And from what I've heard from you, it also some of them found it easier actually to communicate and to learn in this way than in a more messy setup with the many other unpredictable humans.

Olli: Exactly. And this is one big area of social robotics where there has been many products created actually to teach autistic or people with autistic spectrum disorders, new things, you know, conversational skills, social skills. And one reason is that when it is a robot, it is a controlled, more controlled environment. So the robot does not convey any unconscious messages that humans do during social interactions. And also, we can kind of limit the amount of different stimulus as we want by designing the situation in a way that is suitable for the person with this disorder.

Silvija: So. There was also another story you told me you have to explain a little bit more on how these social robotics can help interaction between different disciplines and fields. 

Olli: Yeah, what is the idea? So that's about how I see the social robots. The kind of field developing is where we will need to have different competencies of people creating these use cases for the robots. So now we have the technology. But I think one big challenge still is knowing how to apply it in different environments. And this is where we also try to kind of have a very cross-disciplinary team at Futurice when we work with robotics to have a have a different domain expertise, whether it's applying it in healthcare or the retail sector or in some other area? It's very important to have the domain expertise converging with the robotics knowledge and interaction.

Silvija: So if you have a receptionist in a hospital or in a doctor's office, that's actually a social robot. They would need to know some medicine, some psychology, some I don’t know to build the right thing.

Olli: Exactly. Exactly. And this is one point I want to make. There is no general purpose robot. There is no generic robot that you can simply put in a place and it can do the tasks that are required. It's always in contact with the environment where it is being applied and to make a successful application. We need to have the integrations made to integrate the robot to the environment and the operations and the system that are actually taking place in the environment that it's being applied to.

Silvija: Very cool. You mentioned there is a narrow and wide area. What would be good examples of narrow and is it so that white is better than narrow or just different?

Olli: So my definition about social robotics, the narrow definition would be that it would be a physical robot that's primary function is to interact with humans. And one example would be a kind of robot whose function is to teach a language to you via speech and dialogue. That would be up this narrow under the narrow definition. And then the wider definition would be, for example, a mobile robot that operates in a hospital and its task is to move a dirty laundry and medicine inside that hospital. But it will, you know, meet humans along the way and it might want to tell them something, why it's there, what its purpose is, possibly even ask them to step aside or if there is something like this. So the new area and are the new wave of robotics that are being now applied in these social environments. They need some kind of layer of social competence or social capabilities for them to be safe, for them to be nice. Nice as in in terms of user experience. So people want and expect robots to understand the social norms. And some basic mechanisms of operating in the spaces that we operate in.

Silvija: For Norway, you mentioned that you like no isolation. 

Olli: Yeah. 

Silvija: Please help me. I love no isolation. But I think of them as a telepresence company. You. You say they are social robotics company. Why? 

Olli: Well, I. I think so in social robots, we always have the actual robot. And it can be autonomous, which means that it actually operates by itself via some kind of a pre-programmed behavior. And it does not need a human to actually step in and operate it. But then we have what we call the Tele operated or telepresence robots, which are a different kind of field, which no isolation is also a part of where we have the robot as actually an interface. And through that interface, a person from, let's say, another part of the world can meet and talk to another person.

Silvija: He can be our ambassador in Norway. 

Olli: Yes. So you can think of the robot as a mediating device. And yes, of course, it matters what kind of a box or whether it's a humanoid or what is that the device that is in mediating this discussion. So, I mean, we have a voice that comes out of the robot. Maybe it could have some lights. Maybe it can even do some motions, because that makes it more lifelike and it makes it more fun to interact with.

Silvija: Very, very cool. Who should we read to learn more about this field?

Olli: I think it's there's some interesting stuff going on in the M.I.T. Media Lab with the research group of Cynthia Brazille. I've been following that that that a lot. And then, of course, if you're interested in Android Development, then Hanson Robotics is one. And then the Japanese professor at Ishi Guro who is developing also androids would be something interesting to follow. And then there are if you want the more technical side of social robots, I think Cynthia Brazille has released some books as well that you might want to take a look at.

Silvija: You're not mentioning Boston dynamics. 

Olli: No.

Silvija: they are not social. Those dogs and those Atlas robots.

Olli: yeah, that's a good question. And I guess they would fall under the category of the wide, wide definition of social robots. So if they are moving in environments among humans opening doors, they would need to understand some basic human behavior to operate safely and nicely in their environments. But their primary function, I see, is not communication.

Silvija: I have to say, one of my eye openers here was when I saw one of the first Boston dynamics robots where there was this atlas who looks like a human sort of. 

Olli: Yeah.

Silvija: And then they were trying to get him to fall over and pushing him with a stick or broomstick or something. And, you know, he was actually quite good at keeping his balance. But at some point, I noticed I was really angry at the man who was trying to get the poor robot to fall over. And, you know, I was thinking, you know, it won't be long before we actually develop quite a lot of empathy for these things.

Olli: That's true. That's true. And then there are other studies where they have looked at. I don't know if these are ethical studies, but they've been looking at the, you know, trying to a little bit like a play around with a hamster and then also looking at what is the human empathy towards that or what reactions that creates. And then also looking at having a simple robot and kind of tossing that around a little bit. And actually the reaction is, is more or less the same between the hamster and the robot. So we feel as bad for both of them, which is very interesting, I think. And it's actually touching that area what you just…

Silvija: …you know, first I thought he was bad at manipulating it, but then I thought, you know what, it's good because that's what humans are. We are empathy machines. Good thing I feel something bad for the robot because you know, otherwise it. 

Olli: Yeah. Yeah. If you wouldn't it. Yeah. In that sense. Yeah.

Silvija: I smiled when I read you I was asking you for your favorite robots quote.

Olli: Yeah.

Silvija: And you were saying, you know, I'm C-3PO human cyborg relations. We also think about C-3PO and R2-D2. And I was wondering if I should ask you about the difference between the two because it's C-3PO. 

Olli: Yeah. 

Silvija: Is this superhuman looking? 

Olli: Yeah. 

Silvija: But completely intelligent emotionally in an unintelligent robot, while little R2-D2 never looked like anything we can, you know, associate with, but we all love him. Probably the most loved robot in human history. Why?

Olli: And that's a very, very good question. I guess it. What we can take away from this, and I don't have the answer for your question, but it would seem that we can create the feeling of empathy and sympathy to other things that are unlike humans as well. So if you look at actually this behavior, it's not a speech that is difficult to humans, but there is something, there is motion, there is blinking lights, there is beeps that kind of create a presence that we start to love and like whoever says, C3PO is more like annoying with its behavior

Silvija: over-rationalization 

Olli: Exactly.

Silvija: I think he flew the communication of R2-D2. He communicates that he understands you and that he cares and he's extremely loyal.

Olli: Yeah. 

Silvija: And I think we all fall in love with that loyalty. 

Olli: True. 

Silvija: And so I think we shouldn't underestimate the power of all these social robots in all this kind of growing loneliness and isolation that the world seems.

Olli: Yeah, definitely. There are there are many possibilities in that field. And also why I chose that quote this because I feel that creating a kind of general purpose social robot is still most of the or everybody's dream. We want to reach the people that are creating the social robots. I think they aim towards that kind of general purpose robot. But it's still a little bit off, you know, so

Silvija: Olli Ohls from Futurices. Thank you so much for coming all the way from Helsinki to teach us about the power of social robots.

Olli: Thank you very much, Silvija. Great to be here. 

Silvija: And thank you for listening.