Ethical implications of In-Home Robots - Guy McDudefella, Brittany Postnikoff

Britney post Nick off so hi this is welcome to my talk so why are you here easy robots so who can name these robots

the last one is meta be from medabots so I grew up with these robots like this is my childhood this is what i did on weekends was i watched movies i read books it was fantastic but what makes these robots memorable well how they interact with things specifically their how they interact with their environment and the people in that environment this is something we can call personality in the same way that this car could be described as stubborn you lend it to a friend you're like oh no no no it's a it just stubborn you know it totally works me know what's [ __ ] get a new car but we have the same sort of

relationship with robots so before I get a 2 into this and I do a bit of my credentials I am a paper collector I've got the two diplomas a degree i'm working on my masters and the best part about this is the experiences I've gotten with it each one of these has taught me how to well manipulate people and then manipulate in specific industries and well it got to the point where i got into robots how do i get robots do this job for me this is mainly in my undergraduate so we'll go through some of the labs they did there the first lab i joined was the autonomous agents laboratory i did a bit more

hardware part of it here and this is my favorite project I developed a robot that kid autonomously downhill and cross-country ski but people love this project and we were like what makes us different than the other projects at the University and while that robot is [ __ ] adorable the factory it had a hat it had a jersey people like we could totally identify with this this is just like a tiny skier well I also saw this when i went to china for robot olympics or hero cup we would have team pride come about so these were our outfits again we had to keep the jayne hat but we also had our own matching jerseys and

as we did our competitions people had actually cheer for us instead of their own team because we had these outfits so like yeah yeah but I can totally cheer for you guys and I was like wait that's really weird so it's just a robot so what else can I do with this oh yeah joke

this was the actual yeah the actual effect so once I figured out that people could feel for robots I decided to join the human-computer interaction laboratory which also has a subdivision through human robot interaction laboratory or HR I shortly after joining these were my two favorite books I'm like wait how can I get a robot to follow these rules or the stuff I can learn in these books and found the first piece was having a body having a face having appendages being able to act in a humanistic manner so i'll be talking about humanoid robots mainly today when we have these robots we like naming them just like people named boats like the name cars like they named while even

their computers sometimes people love naming things and this happens with the robots as well people also maintain their robots they're very big about well if you bore this robot it better come back in perfect shape this is my robot this is my divisions robot and they get this possessiveness they also assume responsibility over the robot if the robot goes missing or if something happens to it they actually get very very upset it becomes a friend in the workplace so I thought like how can we exploit this and I decided to target empathy so the first thing we did Oh empathy so it's basically what Pixar does in every single one of their movies they make us feel for things right like

that movie um but can I make this happen with this adorable robot like it has big eyes just like you know baby animals it's only about this big it's about 10 pounds but couldn't make people feel empathy for this robot what we did was we had this robot sit across a sudoku board from a person a participant in an experiment and we had them build a rapport like how are you today how's the weather and people would say do you even know it weather is hahaha like everybody but it it built rapport because they were they could ask these questions in the robot would answer they could have a joke together shortly through the experiment

like 5-10 minutes in after you could do a little bit of rapport we had the robot say I have a virus i think i'm sick and i would start jittering and start hitting itself it was very obviously something was wrong with it and then it said I'm afraid the researcher is going to reset me so what do we do right well we've reset it and people got upset they had their they were like oh no like and they and they actually would cover their face with their hands they would like they were actually visibly upset and they would shake and it's like that's really cool what else could we do but then I was like how can this be applied

right so this robot here what it does is it picks people up off of battlefields and carries them back so somebody gets shot or injured this robot will go and save them this is helpful because then it saves people from you know spending another life to try and save somebody which is great and like I mean it's nice when can we can save people but imagine this robot goes too far or it gets in a hole and the other side comes across it and booby traps it well we talked about the names we talked about the empathy we talked about how people assume responsibility for the robots so what happens when somebody goes to save the

robot and they get um blown up shot something bad well it's kind of doing the opposite of what it's supposed to do at that point so empathy can also be used against us so my next thing was Authority so how do we get robots to have authority over people and authority is usually done by making people do things they really don't want to do like rename a thousand files by hand so we did this in an experiment we had people start with 50 files a hundred files 500 files and one guy even got to a thousand files like he renamed all those like f to type something f to type something and it was like ours

so we have a robot versus I came in who do you think is going to win who do you think has more authority robot human robot so what had to happen is the person had to say no three times before they could get out of the experiment so little Milgram II like we had to jump through some hoops for ethics boards for this type of work um but the robot could win and people swear at it they would get mad at it they would have liked Oh they'd have their outbursts something they wouldn't do in front of the person this part we need to study a little bit more but the the numbers are right like

the robot made people do more so now let's think of abuse of authority right this is a robot that you can use to dispense medication in hospitals so what it does is it make sure it makes sure you stay live I'm pretty straightforward but what if this robot was compromised what if this robot looked like our robot we were talking about before what if the robot could do multiple jobs so you go from something that dispenses medication to something that also tells you your daily schedule where when and where you should be and things like that so if this was vulnerable and somebody took control over it well what could they make you do in a regular basis if you

knew if they knew this robot had authority over you and then this is my research project for the last year is robot persuasiveness how could robots to be used to persuade people persuasion starts with argument you have to have different differing opinions for to be persuaded from one idea to another it's just a basic and I wanted to simulate this how do we do this so what I did was I took two human faces like this there's an emotion word at the top and ask people who thinks face a shows joy better anyone who thinks face b shows joy better okay that's actually the most lopsided I've ever seen this the answer is they're actually mathematically the

same so the interesting thing is we had we gave the robot this script um I know it's really hard to read but each of the dots the robot would disagree we had 18 these faces and agree look at the next one agree look at the next one degree then I get to point where you see these triangles and then had three strikes to convince you and this is a script it says to disagree so we'd hit the disagree button we had suggests Oh suggest the opposite face the person wanted and give a reason which was randomized and the same for every single person so it's randomized before the experiment well it still apply to these

faces the robots following a script by got quotes like this they're her cases were now the robot caused me to doubt my decision it was following a [ __ ] script like there's no human aspect to it the the phrases cuz I'm an aggressive person my advisor is like oh we can't use those that those are too convincing they're there too in your face so I actually read a mediation book I read a whole bunch of linguistics books to make sure the quotes I use are very very normal simple nothing to them but it still caused people to doubt their decision they I also got quotes like this sometimes I convinced him the robot also sometimes he told me something I

never realized the robot told them something they never realized it's still a [ __ ] script so and then I got other interesting things like it's it's very interesting to see now look think make decisions and have discussions so people liked interacting with robot lately they liked the experience of being in a room with a robot so now what happens when we have this robot in a customer service role well this robot is used in banks in Japan one bank they're trying this right now and this robot is a teller this is the same robot we've been talking about that the experiments have been done on so this robot can book you appointments it can tell you where your where your

appointment is but that means it has access to your information has access to your account to book the appointment with a loans officer mortgage ER on it's got access to your financial information well that's really scary now when we think about persuasion well what if this robot started up selling would it be more effective than people in the same role would people judge it less because it's a robot do how do we think about morality and ethics with this robot when it comes to dealing with your money and trying to sell you master cards every time you come in or visas or whatever so the other thing too is a lot of these experiments had to go under ethic ethics

boards it was hard it also meant it restricted what we had to do so I can't really test how vulnerable they were in the role I had in human-robot interaction there wasn't as much security or so he'd guide me in that but I wouldn't cover what I found for documentation so this robot again yells out its IP address when you turn it on like actually physically hi my name is now and it will it will do that so the best part was is we had the door open to the room one day and I could hear it down the hall like I was two classrooms over and I could hear this thing yell yelling out its IP address so how can we

use this well it gives you a web page for the default password and username is now and now and then it's done and then you're in so the great part about this web page is you click on a you can type something in and great you can control what the robots saying you don't know how to you don't even have to know how to control the robot you don't have to care where it is but you can type anything you want be is boring see is what you see you can change the password you can change the robot's name I'm d lets you change all of the network connections with one simple button and same thing with e you can upload files

to it you can update it like it's you get all this access when it yells out its IP address and you write it down it's not that hard so this is scary when you think about it being in banks and hotels hospitals so now I'm going to go through a few case studies and examples based on what I've talked about so far this is at this point feel free to interrupt your raise your hand if you have any questions but I'm just going to keep going through until my times up so everyone knows this quote everybody probably has a good idea of ideas behind it things pop into your mind so we'll just move on on tella operated humanoid

robots so the thing with these robots is most of them can be autonomous or they can be controlled so you have sort of two options but how do you know when it switches right we don't have any indicator right now it for example if you have that robot I've been showing all along and so he does start typing sentences in or does start trying to control its movements or adjust things in it you don't know but these robots are actually used in really cool places like at conferences I'm not this conference obviously but other conferences will let you have the rolly robots with the iPads on them that you can use to attend conferences so what if

somebody else decided to overlay your voice or take video of you and edit it and continue to use this thing around a conference how would that affect your image how would that affect how people interact with you in the future yeah you wouldn't necessarily know what happens if you logged on and you're done with your day and somebody goes back and does that this is one of the Japanese robot hotels it only has robots the whole idea is that there's no people involved so you can check in with a raptor even check in with something that's actually fairly accurate to a person or slightly further in this photo they never show it because it's boring is again the same

robot we've been talking about so all three of these are options now again when you think about hotels and how they try and upsell you every time you come in or sometimes hotels are in disaster areas what happens when you need a hotel because something's flooded or you need to get out of your space you don't have your money you don't have things with you normally people would be compassionate they would have an opportunity to say you know what it's on us we understand but how is a robot designed to handle these things how does it affect people who are in a disaster situation and this is a social problem because it's society says well if

somebody is like having a very very bad day and there's earthquakes and tsunamis well we must probably help them in home privacy I notice the original in the title of the talk but things like in home privacy so when you get I mean when you have a child you get used to it you love it even if it [ __ ] up a bunch does this happen with the robot so part of this is is when you get the robot at first you might be a little skeptical you might be like i don't know how i feel about this thing going around my house but once it starts become useful like cleaning up after you or cooking

for you or you know turning down the thermostat if you forget to do it that day because these things like I mean we take it to robot Olympics herb can be somewhat nimble but what if it starts walking into your children's room at night and is recording well how do you know again we don't know when it Stella operated or not so we don't know if somebody has control over it while it's doing things like that or what if it walked into the washroom on you and you're like oh robot like it's a little more advanced in a room but it does have a camera that and it is internet connected so these robots have the same

issues as IOT if you get in through a network you basically a full access to robot you can use that web page you like we're done so again in home privacy if you have bills lying around and all of your personal information out and the robots supposed to be dusting or something and it sees all of your information and somebody's just like piggybacking on the robot well they'll get all your information they'll be able to get to your cows like usually don't think about putting your bills away from your desk if you don't have people coming over me that a robot companion would be ideal or say an elderly person and you put the robot in the house it's cleaning it's

tending it's providing some you know some orderly type functions but then if I was a company I would love to be able access that robot because like you said it can upsell yeah hey have you considered a reverse mortgage as its feeding grandma Jones exactly it's like it's like you think of it to make that ad and he may convince her that she needs to do a reverse mortgage with my company exactly it's like futurama that one part where everybody knows the meme shut up and take my money but that comes from they put dreams and fries head to go by super sexy underwear um that was it they got in through you know a way

that you wouldn't normally think he wasn't prepared for the fact that you get ads in your dreams so this is the same thing you're not prepared like helpful and kind lovely they seemed so right and I learned from this robot now reverse mortgage take out some more insurance by that exactly so this is something we need to be aware of is the social engineering of robots and sort of preparing yourself for it the next one we're going to bring up is robots acting as a security guard so some of this is really helpful because if you think about hospitals they're usually overstuffed you don't need nurses standing at door saying you can you like please not go in here on it's nice to

have a robot to be able to do that stuff but maybe also pick up patients or do this sort of thing so these robots though how do we make them in a way that they'll be sympathetic to people like you're still in a hospital these people might have damaged bodies like you can't react badly like even if you jo baat just stands there we're running against it might hurt a lot right or trying to push it might hurt somebody so this is another social aspect we always have to consider is how the robots are developed what shape does it have is it approachable do you want to interact with it and going back to old folks home

I love this thing it's like this big it's super fuzzy and the idea is same as Paro and it's supposed to be immediate mediating robot or comfort robots so instead of pet theory therapy for elders you can just have this it's hypoallergenic you don't have to worry about it dying except for the power part which you can fix you don't have to feed it you don't have to clean up after it great part about this though is grandparents loved putting sweaters on things like have you ever gone to grandparents like are you cold are you hungry like that that's normal um you put a blanket on the signal like a heavy blanket it can light on fire

it's in the manual like please don't put a blanket on this and it's like you're putting this in an old folks home and saying don't wrap it up yeah but it's still a social interaction like you meet when we design these robots we have to think how people are going to interact with it especially like it's sad but people in old folks homes or when you get older and age might have Alzheimer's or memory problems so even if we try and tell them more things like that you end up having to have somebody seen there and watching anyway to make sure this doesn't happen but seriously if you have time read the manual on this there are

comics everywhere in it about like don't put the seal in water because you'll get electrocuted and it's a super happy looking seal it's getting electrocuted it was super fun to read I chuckled the whole time because of the warning images um but the other fun thing about the seal is that when you pet it and you say word over and over it starts to purr when it hears that word and then it progresses and gets the point where you leave the robot like in a different room and it will still / when i hears that work word so if Graham is talking about global nuclear warfare information leak if it starts hissing or if it starts

purring when somebody else comes the room and grandma says it so this is another social thing you wouldn't think about is you know how is information getting released from petting a furry seal and of course we have robot mediators so robots acting in between people as well same thing if you've out a robot looks very approachable and people want to give it information how does that use how does it use the information in a way that wouldn't reveal other things how does it act towards you because we could these are things that we are already learning to program send me autonomous surgery robots same thing as hell operated robots I'm going to take that out we've also got these robots they're

super cute they pick people up and take them to the rooms I know right but look how happy she is it's like delivery what should we see a lot of these robots in a medical environment and then he also had the one that was in the bank do we have was that now would say that in America would they be subject to like pci or HIPAA regulations I'm Canadian I don't know okay well the one we use a microphone the question is would would would robots have to be subject to pci and HIPAA regulations and you're basically regulations that would control the use of the personal and the PII so this is something I really want to work

with I am the cavalry on because they're doing great things with the medical scene and devices and like i said before on these robots to have the same issues same IOT issues like as soon as you hook it up to like internet stuff happens but the other fun thing is a lot of these have open USB ports on them just like yeah you could walk up and I've got a medical robot whoo um so I would like to talk to I am the cavalry more about this they do a lot of great work but um yeah I don't know but I would love to answer that question for you so come get my card after

oh so okay I'm gonna go back to this one um so there's an actually super great paper that came out of university of washington i loved reading this thing and they actually managed to ddos like a surgery robot in the middle of surgery by sending i think of a single packets like tiny tiny packets and it's like the robot didn't know what to do and it was amazing it's like wow we could totally this is easy somebody's already done this with something they were already trying to ship out the other thing too is this the surgery robot they manage to other things as well like reorder commands or delay commands which is really bad in surgery like yeah when you

think about trying to tie up some of these like veins that's terrifying how did they get past the Ethics Board to let them screw up a surgery that sounds like it would kill someone um so they're fake they're fake body it's like even when you learned even when you learn surgeries as far as i know there's like fake bodies you can get you can and you'd feel in yeah well actual dead bodies as well and yeah so there's lots of other ways to try it and i forget what they used in the paper or if I even read that part of it abstracted them um so but that that's one of the interesting things is there are ways

around or making ethics happy you just don't use it on a live human usually or pets or yeah um sorry I'm not allowed to talk about that um so yeah that's about all i have to say um thank you very much does anybody else have questions I would love love love more questions so like anything so as a hobby roboticist um I can tell you the network implications are immense here and that the ability to d das a robot is generally of the trivial so what if that happens the middle of an interaction like at the bank or something right like it doesn't matter I mean that's the perfect thing but the real danger to me

is not the dosing of it it's the now tele operating of it in a way that nobody knows the state has changed exactly good job on the talk I thought it was very engaging and intriguing one of the things that I was thinking about as you were talking about this is there's like the misuse of robots in order to somehow like victimizer you know you know exercise of vulnerability on a human I was thinking also like as far as improving processes from a business standpoint and then there's some ethical implications for that say say for example like like the person that runs the front desk of a hotel and there's like a preponderance for certain

personality types to argue with that person because the person the human on the other side of the interaction is subjective to things like emotional appeal or you know you know you can say hey I've been having a rough day right so so using I'm wondering if there's research on the use of robots like the examples you gave in order to to try to combat that argumentative like emotional appeal that people will have with other humans and so so trying to optimize the amount of times that I deviate from a process by using a robot on the other end that right that's actually super research that's what other people in my lab have done on other papers we've read

is that's a huge use like we love looking at how robots work together our robot and keep up robots and people work together this one here is called Baxter it's about yay high it's pretty big like Jason Street was under its shoulder when it gave it an awkward hug but this one works side-by-side with people we've used this for quality assurance so the robot can actually make up or the person can make a piece the robot will copy and it will be like I think this was bad do it again and so and so that's part of it is when somebody can't do something on a factory line properly they've been using these one of my fellow researchers was

doing research into using these robots be like nope do it again nope do it again for things that couldn't do itself so yeah we like working side by side we like keeping people on a path and yeah production lines are one place we need to keep people on a path

so very related to the last question but you talked a lot about human robot interaction are you aware of any research in the field of machine-to-machine communication or robot robot communication and how I might assure trust in in robot mediated systems on I only did that a little bit with robot soccer and it didn't usually end well it usually end with like one robot like people's elbow into the other robot during a soccer match so no not really seriously watch robot soccer videos it's pretty much just like dogpiling constantly they don't mediate well okay on the social engineering research or social engineering vibe robots the authoritative pneus and the persuasiveness did your experiments consider the perceived gender of the

robots yes so um that's one of the problems with on the white robot that I've been talking about mainly is like look at those biceps like I'm in an I know um like wait that one's really good you can see like it's got molded biceps it doesn't need that like it was who's it trying to impress

exactly it is very much a male robot um and yes it is AB robot we joke about that all the time because it looks like a little football player it's like this guy is stalky he's rather he's dense for his size but I'm trying to find this other so this other robot this one's very gender-neutral and this is when I used in my other lab and I loved it very much because I was living with the cousin at the time I need a seven-year-old daughter brought this home and she's like what's her name and I'm like [ __ ] yes um and then we named we had for them Jose Jennifer [ __ ] I

forget the rest because they only cared about to but I'd brought Jose home a bunch because he has certain programs may have Jennifer do other things because we changed a little bit but I brought Jennifer home the day I'm like oh because my cousin's like no it's a boy I'm like no this one the girl so it does affect things but unfortunately the labs don't share the resources well so I did most of my experience with the male looking one on so it's something I want to do in the next iteration which I'm working on right now but you're right it does have an effect I just don't know on this level persuasion and I really

really want to get into that so thanks for the question and I will hopefully follow up with you in a year any more questions is there any research or you know are the companies being held liable when the vulnerabilities are found you know I just wonder where the responsibility falls is it on the user isn't on both you know we're so I'm sorry um it is a very good question though like as I was going through these robots I only found one that publicly posted they had a chief privacy officer otherwise there is no real security part to any of these and I mean they you can look I mean most the code is open the black ones the

Darwins we can do anything we want those no problem the nails are very very very controlled like is really cool we tried to reset the operating system on it because they come off a warranty and it's like time to experiment um we tried to install a new operating system and we left it for an hour and I Brittany its operating system with the old one so it was like oh ok we can't do what we want with it we like if there any vulnerabilities we even fix ourselves like there's not really an opportunity to make it better but yeah definitely watch the other talk it was really good

frustration with robots I'm a crotchety old man and I could see myself getting very angry because it doesn't hear what I'm saying properly and it or it's it's not you know Here I am having a rough day and it's not giving me a special treatment I want out what what happens when people become frustrated with them do they like lash out and hit do they storm away do they I have a great ride okay great so back to the persuasion one um because it was that script it was a script I'm actually a little bit lazy I didn't get the timing are I didn't have the time to sit there and make sure the robot could respond on time whatever so

it was till it but just to the point where it was only pressing buttons there was no other human interaction or human control of the robot but so my fellow researcher thought was it is he was sitting there pressing the buttons quite literally on the robot and I would actually make people mad we had one person be like no you're a robot don't you dare like talk back to me and they got really Reggie um and back to ethics for it's like I my fellow researcher was like I actually feel really uncomfortable I've never done that to somebody before I've never made them this angry and so there's something we want to research more but yeah like we

need to have these sort of protocols in we have to think about how we're affecting people like when people get so angry that's when you have really bad like social events happen is when one group gets angry enough I hope it's not the rising of the seniors that might be extra terrifying

but yeah like aggravation is a real part of dealing with robots it is so social robotics isn't a super old field obviously because mom computers first you work your way up but yeah that's definitely something somebody needs to research into so I'll try and see what I can do so here's an additional thought I inherently social engineer things robots AI auto callers political pollsters it's fun I have so many papers for you do you run into that or have you considered that aspect so um I just actually graduated the undergrad parts so I it's all been in labs and like I said the ethics boards it's very hard to get anything outside to try and like ethics

pores are so hard um but one of the parts about it is it has to be in a controlled space I can't let just anybody walk by and if I tell you to a social engineer it like most of what we do is suspend belief right like it's just the fact that you believe it's doing what it's supposed to be doing that makes it work the same way we think magicians right like you assume like oh my god that guy's magic of course that's amazing like I went to watch her show last night and it's like yeah I had my magic friend with me he's like I'll watch that thing watch this thing and I

go but I can totally see what's happening still a great show because I'm like I get to learn things but it's a suspense belief or disbelief that makes things like that work and that's the same way we do with robots actually when we have a second researcher controlling the robot we call it wizard of oz' ng

and then we'll be equal I think will be equal because we treat robots the same way like it's yeah the great part about these types of robots are though is that they have a body and the fact that they are doing gestures meeting your eyes they they get the full social training but the best part is they don't have bad days they don't they're not there forget their training like as soon as I showed some of my friends friends this that work in other companies like we can replace that girl they can't get her stuff done every day because they could have this robot that won't be sick it won't take extra days off or the guy

that like won't finish is filing every day because they just do what they're supposed to do and they act how they're supposed to act that's why we named them that's why like people will like do events with these robots that's why they work in customer service is because we treat them like people and this is why this whole talk is here so that's why I think it's going to be equal because there's a body and I can give you a video on this as well where one of my researchers that was part of the empathy thing was his full experiment was on whether we empathize more with on-screen robots or in-person robots and in-person robots was way more effective so I think

it'll be about equal the people and humidity and robots another question so I'm wondering if being this is like you know somewhat of a hacker conference you know going down the route of using a robot to pivot and maybe get information that you couldn't get in person I'm imagining like I'd be less likely to tell someone random a person like my private Wi-Fi password at my house but if a robot that was on my network came up to me and asked me I'd be more likely to I think less about that decision like okay well the robot probably needs it so in that vein or like in that topic I wonder if there's research that could

be done related to what type of tasks are people more likely to be manipulated by a robot than a human social engineering engineering tasks are like clicking links and a variety of things that happen inside of a social engineering space that it's interesting to see well what's you which what types of those tasks are easier for a robot to accomplish than a human when they're doing social engineering the first thing that popped up in into my mind is those annoying credit card people at airports I I could see more people going to a robot than somebody who's like hi I can get you get you to stop besides this adorable roll up being hi can I give you

a credit card I'd probably talked to the robot I don't talk to the people so I could see that being more effective persuading you to get more airport master cards or visas I guess I should be saying credit cards Wow product placement but that's that's part of it like that's a really good question is what tasks would they be better at and I think it's things we already discount people for like you like who answers the phone surveys anyway like yeah exactly but but it's something where as soon as somebody picks up your leg off [ __ ] and you hang up well robocalls didn't work there but again these annoying sales people like in the middle of the mall

trying to give you perfume it's a cute robot it's different I might start speaking to it whereas the people you already know oh this is something I can ignore it and walk away so that's the thing though it's levels of adoption so it's where you start and right now we're at the forefront of this so it's something we can know now yeah so that was sort of the purpose yeah um I guess I wonder if there have been any studies for many of the organizations that have like like when I call my bank and they have an automated tell you know system if they're actually monitoring any of my verbiage when I'm getting frustrated to

recognize like we're just going to send her straight to the representative she's getting really pissed off because apparently what she's saying isn't being recognized we're going to we're going to recognize that I wonder if they've done any studies that you could resource for the empathy you know pull like something where you could recognize okay they've already they've already noticed certain vocal recognition or certain things that people say or do that might cause a heightened sense of frustration that you could resource to try and make sure that the robots react accordingly so the fun part about all my research is I haven't had to touch a I at all most of what we do we want to look at how having a body

changes things and how the like gesturing and gaze is like almost over done already because people keep like I gaze is so important like if I was looking over here and responding to you like you would feel unappreciated so that's what we test the robots is if we have that same effect on people AI is a whole different lab that I tried for a little while and didn't want to stay but that's sort of the thing with social robotics is a lot of it is just was er de passe it because we have other people working on that just like I don't do any computer vision with these because we have other people to do that but i would

probably source other people to do that for me and figure that out but when it comes to social implications of that I'm much more interested so I'd probably just fake something together yeah quick question so fellow Canadian represent whoa um so I'm a big kind of video games and one of the things they've been exploring in video games with robotics particularly the gas in mass effect and the aam Knicks in blizzards overwatch is they've been exploring these worlds where we're robots and make them completely self-aware and now they're fighting for their own rights to survive and be part of the rest of the world and we treated equally in your opinion did you see that as an actual feasible

future that we could one day find ourselves in and if so what's what type of barriers what after we overcome to get there I actually heard a great title this week what was it it was a robot anthropologist or something like somebody's actually already got this job title that they're studying like the culture of robots like the culture that robots have on their own it's like what that is so cool like I want this job but that's what there are trying to do is have you watched futurama and you know how they always have like Rollo kind rallies and benders always like kill all humans except one one great thing about robots is they're very easy to figure out what social

interactions we need because we already have pop culture like we already respond they expect robots to be able to do what they can read in books what they can see on TV what they can see movies so anytime i bring participants in like it's like in the movies like or it's like that book i read and they get really excited because they already know how to interact with it so when you're talking about yes do I think this is going to be a thing do I think Robo kind is going to be a thing I think is the gist of your question I think yes because we already see it in pop culture so we're going to want it to be but

that's like holy [ __ ] ways out all right we've got time for one more I know when a robot is assembled it the more it looks like a human the closer approaches the uncanny valley is there also a verbal uncanny valley if it's scripted responses I know okay I'm dealing with a robot but if it's a very interactive response i can tell i can tell the two between human and robot conversation but is there a point at which the robot conversation is just too close but not quite and it's like whoa crap get me out of here so this is very canadian moment on the robot trying to get it to say face a or phrase be but it would say fuh

say instead of face a if I typed it as face a so the only way to get this to work was do facespace canadian eh so it was close enough but we did have to do some fudging to make it not reach verbal uncanny valley so that's a very valid question and it does happen like I mean also like this the sophistication of the communication I think you should choose be well I want it really I think you're being irrational should you be um we straight away from that because we get too much human element in and we'd get too much of you know if I was sitting there typing in then it's way too much of my opinion and it's not

testing what we're wanting to test um but I'm sure hobbyist maybe you want to play with that part

that's a real big hint that you know you are really listening to me and sometimes that means you're a real person

detecting his description is whether how is being delivered is not know yes I'm happy the program the bioware / non-fire they're really imagery of that yep so that's great thank you everybody for coming give me questions