Latest / Automated with Brian Heater / Kate Darling on Why Humans Bond With Robots

Transcript

Kate Darling on Why Humans Bond With Robots · Automated with Brian Heater

0:00
SPEAKER_01: I'm just like so convinced that every social robot we've seen so far has been like the palm pilot of innovation.
0:07
SPEAKER_01: We're so skilled at creating genuine relationships with all types of things and people and animals.
0:14
SPEAKER_01: And I think robots are just like a natural extension of that.
0:19
SPEAKER_00: Part of the criticism or concern there is that maybe our empathy or our emotions are like a finite resource.
0:26
SPEAKER_00: And if we put them onto this robot or put them onto this chat bot, then maybe we don't have enough left over for human beings and animals.
0:33
SPEAKER_01: I think our attention might be a finite resource.
0:37
SPEAKER_01: So I'm worried about the ways that technologies get designed to manipulate people's attention and time and money.
0:45
SPEAKER_01: The whole promise of humanoids is that they can be in spaces that are built for humans and probably still have humans in them as well.
0:52
SPEAKER_01: So it's really important to, you know, get the safety right for the technology.
0:56
SPEAKER_00: We're reading a lot of these like horror stories, you know, certainly like people doing self-harm and things like that.
1:02
SPEAKER_00: So the question ultimately is like how and can we hold these companies accountable when things go bad.
1:11
SPEAKER_01: Let me back up a little.
1:13
SPEAKER_00: Um welcome to Automated.
1:27
SPEAKER_00: My name is Brian Heater.
1:28
SPEAKER_00: I am the managing editor at the Association for Advancing Automation.
1:32
SPEAKER_00: Please forgive my voice during this intro.
1:34
SPEAKER_00: I'm currently battling the Parisian flu.
1:37
SPEAKER_00: This, however, is a very fun interview we did with Kate Darling of the RAI Institute.
1:42
SPEAKER_00: Kate brings a unique sociological bent to the Institute's work.
1:47
SPEAKER_00: We also discuss my favorite story from her latest book involving uh the ritualistic sacrifice of a Plia robot.
1:54
SPEAKER_00: It's it's fun for everybody involved who isn't a robot dinosaur.
1:58
SPEAKER_00: Uh thanks to Kate and RAI.
2:00
SPEAKER_00: Thanks to you as always for tuning in.
2:01
SPEAKER_00: Don't forget to like and subscribe and check out our newsletter over at automated.fm.
2:06
SPEAKER_00: In the meantime, please enjoy this conversation with Kate Darling.
2:11
SPEAKER_00: Vision and AI are changing what's possible in industrial automation from quality inspection and predictive maintenance to robotics, logistics, medical devices, aerospace, and manufacturing.
2:22
SPEAKER_00: At the Advanced Vision and AI conference, you'll hear from industry leaders, see real-world AI applications in action, explore the latest in machine vision, computer vision, sensors, edge computing, and automation technology.
2:35
SPEAKER_00: You'll also be able to connect with the people who are shaping what comes next.
2:39
SPEAKER_00: Whether you're building, buying, or implementing these systems, this is a chance to stay ahead of the curve and see how vision and AI are moving from promise to practical impact.
2:49
SPEAKER_00: Learn more and make sure to buy your ticket early at the Advanced Vision and AI conference September 23rd and 24th in Santa Clara.
3:00
SPEAKER_00: You know, you and I were supposed to speak uh about a month or two ago, but I'm glad.
3:04
SPEAKER_00: I'm glad that this is happening now.
3:06
SPEAKER_00: And I'm glad that we unintentionally waited for the launch of Familiar Machines and Magic to have this conversation because I feel like this is very relevant to the work that you do and that they are essentially launching a the latest version of the robot pet.
3:27
SPEAKER_00: And I wanted to get your first unfiltered impressions on the work they've been doing.
3:35
SPEAKER_01: Yeah, I have to say I'm super biased about that particular project because I talked to Colin about it when he was still just like ideating shortly after he left irobot.
3:46
SPEAKER_01: And I remember like we were in a car at some point for two and a half hours with like two other people driving to this conference, and he and I were just like totally vibing on like how you know there's been this social robot winter, and whoever figures out the killer application in social robotics is gonna like really make a difference.
4:10
SPEAKER_01: And just I I'm still I'm just like so convinced that every social robot we've seen so far has been like the palm pilot of innovation, right?
4:18
SPEAKER_01: And like at some point, someone is gonna get that robot into enough people's homes that people start to genuinely appreciate the value, the social value of having it around, not just like it has to have some functionality.
4:31
SPEAKER_01: Um, so I don't know.
4:33
SPEAKER_01: I yeah, I've and I've been in touch with Colin throughout like this whole project.
4:39
SPEAKER_01: I think they're doing a lot of things right.
4:41
SPEAKER_01: Um, I love that it the robot doesn't speak.
4:44
SPEAKER_01: I love that it's mobile, I love that, you know, it's somewhat privacy preserving and that like all of the stuff, you know, runs on board.
4:53
SPEAKER_01: So yeah, I, you know, I'm a fan.
4:55
SPEAKER_01: I don't know if it's gonna work because a lot of it depends not on the robot itself, but on whether it can get enough adoption that people really understand the value of having it.
5:05
SPEAKER_01: Same, same as the palm pilot and the iPhone.
5:08
SPEAKER_00: Yeah, uh you you said that you know the issue is that um nobody had quite, I guess, unlocked the killer application, which implies that like there is a killer application for a social robot.
5:20
SPEAKER_00: Um and this is work that you've been doing for a really long time.
5:23
SPEAKER_00: I mean, do you have like a general broad sense of what that might be?
5:27
SPEAKER_01: Oh, the application is social emotional.
5:30
SPEAKER_01: It's like I I mean, pets used to have to house some kind of functionality.
5:37
SPEAKER_01: Like dogs would guard your house and my the cats would catch the mice.
5:41
SPEAKER_01: Um, and then once they started being part of people's families, people realized, oh, actually the real value here is the companionship.
5:49
SPEAKER_01: And I don't think people think of people are starting to, I guess, realize, especially with like more advanced chatbots, that there is a real companionship aspect to a lot of these technologies that is you know very valuable.
6:03
SPEAKER_01: Um, but when I first started working on this, like that was still like a crazy idea to people that you know we would socially, emotionally bond with a robot and that that would be, you know, it made people upset the idea, oh, like that's gonna replace human relationships.
6:18
SPEAKER_01: And, you know, I I just I think that there's we're such relational creatures and we're so skilled at creating genuine relationships with all types of things and people and animals.
6:33
SPEAKER_01: And I think robots are just like a natural extension of that.
6:37
SPEAKER_01: So for me, there's been no question that there's gonna be like an immense value, just a social value of it.
6:44
SPEAKER_00: Yeah, I wonder if part of the um part of the criticism or concern there is that maybe uh our ability to socialize, or maybe if like our empathy or our emotions are like a finite resource.
6:58
SPEAKER_00: And if we put them onto this robot or put them onto this chat bot, then maybe we don't have enough left over for human beings and animals.
7:06
SPEAKER_01: I think our attention might be a finite resource.
7:09
SPEAKER_01: Um, so I'm worried about the ways that technologies get designed to kind of manipulate people's attention and time and money.
7:21
SPEAKER_01: Um, but I think when it comes to just empathy or how we relate to others, I think that we have incredible capacity.
7:30
SPEAKER_01: I don't think it's a zero-sum game.
7:31
SPEAKER_01: Like I have two kids.
7:32
SPEAKER_01: When I had my second kid, it wasn't like I love the first one half as much.
7:35
SPEAKER_01: Like that's not how empathy works.
7:38
SPEAKER_01: Um, in fact, there's research showing that, you know, people who care a lot about other entities, whether it's spacer or animals or other people, um, it's it it just tends to kind of multiply rather than that taking away from you know their empathy in other places.
7:58
SPEAKER_00: Yeah, I I I become this is not a a new observation, uh, but I become increasingly convinced that the root of like the vast majority of our social ills right now is this this isolation, this increasing isolation that we've had.
8:13
SPEAKER_00: And, you know, I I wonder if this is to what degree some of these things are a solution to that issue, and to what degree they might further be isolating us from one another.
8:28
SPEAKER_01: I think that depends entirely on how they're designed and for what purpose.
8:33
SPEAKER_01: Um, if you look at again, pets, pets have been just an amazing value add for people and in helping them deal with loneliness.
8:43
SPEAKER_01: They're not a replacement for people, but like if Uncle Jerry gets a dog, we're not like, oh no, this is gonna, you know, take away from his human relationships.
8:53
SPEAKER_01: We're like, this is great.
8:54
SPEAKER_01: This dog is gonna help Uncle Jerry, right?
8:56
SPEAKER_01: So I think there's a way to frame these technologies that are really like to the user's benefit.
9:02
SPEAKER_01: I think the problem that we are already seeing and gonna see a lot more, is that the business models tend to be a little bit different and the technology tends to be designed, you know, in a company's best interest and not a user's best interest.
9:20
SPEAKER_01: And that can often involve taking away um from existing relationships by monopolizing attention or trying to manipulate the person into certain purchasing patterns or or what have you.
9:36
SPEAKER_01: So I think that's the issue.
9:37
SPEAKER_01: It's not the technology itself.
9:39
SPEAKER_01: It can be designed in a way that like fully supports people and their relationships.
9:45
SPEAKER_00: Do you have faith though?
9:46
SPEAKER_00: I mean, you know, I I think you know, we maybe are similarly wired to be cynical against certain things, um, you know, and and and when it comes to, I guess, certain certain interests.
9:57
SPEAKER_00: So do you have faith that a product like this from a company like this can be um, you know, both I guess uh successful but also socially beneficial?
10:09
SPEAKER_01: I think it could.
10:10
SPEAKER_01: Like, I think the tricky part right now is that there's no market, there's no pet market for robots.
10:18
SPEAKER_01: But I think that's exactly what like a company like this is try is trying to solve for, right?
10:25
SPEAKER_01: Like they do see loneliness as the application, and they do see that at some point, and I truly believe this, people will pay just to like have the companionship without needing to have you know some surveillance aspect or like it guards your home or you know, what whatever it is.
10:44
SPEAKER_01: Um I just don't think we're quite there yet.
10:48
SPEAKER_01: We're I don't think people have realized yet that like the willingness to pay just still isn't there.
10:55
SPEAKER_01: So it'll be really interesting to see if this company succeeds in creating that market or whether they have to fall back on some business model that is ultimately not in the user's best interest.
11:07
SPEAKER_00: This is a really interesting point, and this is something uh, you know, Colin and I I spoke about this ahead of the launch.
11:12
SPEAKER_00: And, you know, if you iRobot has a really interesting history that I'm sure you're more aware of than than most people.
11:20
SPEAKER_00: Um, they they tried a bunch of things out, they had success with the Roomba, and then most of the attempts to kind of recapture that success were similar in that, like, let's launch another home robot that can that can do something.
11:34
SPEAKER_00: And none of them really managed to capture that success.
11:38
SPEAKER_00: So this is the question that he's getting a lot right now is like, okay, yeah, this is interesting, but what does this product actually do?
11:45
SPEAKER_00: And this is like an uphill battle for him right now.
11:47
SPEAKER_00: He's is he has to say, well, it doesn't really, you know, it's just it's there.
11:50
SPEAKER_00: It's you know, it's a it's it's a companion robot.
11:54
SPEAKER_00: And like, what's the value add, I think, is what people are asking it right now.
11:59
SPEAKER_01: Yeah, that's the that's the tough part.
12:01
SPEAKER_01: And I think there is a value add, but I think people don't see it yet.
12:03
SPEAKER_01: Yeah.
12:04
SPEAKER_01: People don't see it yet.
12:05
SPEAKER_01: So the question is when when will we be there?
12:09
SPEAKER_01: There have there has to be enough adoption for people to see that value.
12:13
SPEAKER_00: Yeah, the the fact that the market doesn't exist though is not for lack of trying, right?
12:18
SPEAKER_00: I mean, you know, you mentioned all of these attempts before, and I know like you are at MIT Media Lab, so you're very familiar with like Jeepow.
12:24
SPEAKER_00: I know that you're very familiar with PLIO.
12:26
SPEAKER_00: We'll get into like your familiar area with PLIO.
12:29
SPEAKER_00: Um, Ibo is like probably the best example of this, you know.
12:32
SPEAKER_00: Sony's tried it several times.
12:34
SPEAKER_00: There was some adoption, you know, and and uh in a lot of ways, I would say Japan is probably maybe ahead of us when it comes to at least like being accepting of some of this automation.
12:45
SPEAKER_00: But this has been tried for decades and decades.
12:48
SPEAKER_00: Um, what's your sense of the missing piece or the missing pieces there?
12:54
SPEAKER_01: Yeah, I think really it's just the same, the same way it's gone with animals.
12:59
SPEAKER_01: Like throughout history, most animals have been tools and products, at least in kind of Western society.
13:05
SPEAKER_01: And then in more recent history, they've become, or some of them have become our companions.
13:10
SPEAKER_01: How did that happen?
13:11
SPEAKER_01: Um, it happened by enough people purchasing them for this like functionality aspect that I just mentioned like dogs guard your house, cats catch the mice.
13:23
SPEAKER_01: Enough people had that close proximity with the animals that they started developing these emotional attachments, and that then became the value.
13:32
SPEAKER_01: And so I think that the same way with robots and and AI as well, like chatbots, I think that if there's enough adoption, people will start to understand that there's a value there that they weren't seeing or believing in previously.
13:51
SPEAKER_01: The question is, can we get there without entrenching the whole industry in other kinds of business models?
13:58
SPEAKER_01: And I don't know.
13:59
SPEAKER_01: I I'm I'm actually pretty hopeful.
14:01
SPEAKER_01: Like you see this already with like advanced chatbots, and a lot of people and people email me like every week and they're like, hey, I never thought this would happen, but I'm feeling really emotionally attached to this chatbot that I'm using.
14:13
SPEAKER_01: Like people didn't do that, you know, three years ago.
14:18
SPEAKER_01: Um, people would still kind of view that as, you know, a niche that someone might, you know, develop an emotional connection to a machine.
14:27
SPEAKER_01: And now it's starting to become like really obvious, not just to users, but also to people in the research space, that, you know, this the fact that people are going to talk to chatbots not just because they can give them an answer to something, but because they're lonely and because they're getting a lot of like social emotional value out of it and like all of the issues that that raises in terms of consumer protection, all of that, like in that space, that's already happening now.
14:53
SPEAKER_01: And it's also in part why I've like kind of moved on from the world of social robotics because I feel like that's coming.
15:00
SPEAKER_01: Um, it was like interesting to work on when it was something that people didn't really see as coming, like beating the drum, being like, no, we see it from all this research.
15:13
SPEAKER_01: People are gonna, you know, project life onto the robots and they're gonna treat them not to be.
15:18
SPEAKER_00: You were into social robots before they were cool, and now that you know, now that they're cool, you've moved on.
15:22
SPEAKER_00: Yeah.
15:23
SPEAKER_01: Exactly.
15:25
SPEAKER_00: So I mean it's it sounds like by and large, like uh assuming, and and obviously this is like a big assumption to make, but assuming that um the uh that the companies are um you know coming from a uh a generally good place that um that these sorts of attachments um can can largely be a net positive for people and society.
15:53
SPEAKER_01: Companies coming from a good place is quite a statement.
15:55
SPEAKER_00: I know, I know, well, I know, I know.
15:57
SPEAKER_00: That's what I'm saying.
15:57
SPEAKER_00: I I know that we have I know that we have I know I know like I'm saying like extracting to zero, like just like the existence in a vacuum.
16:07
SPEAKER_01: I think absolutely there's so much potential for this technology to be helpful to people and human flourishing and the world.
16:15
SPEAKER_01: Um whether that can happen in the current political economy and the incentives that are being set for companies, that's a different question.
16:26
SPEAKER_00: Yeah, you made an interesting point.
16:28
SPEAKER_00: I was reading, um, I think this was an appearance you made on stage fairly recently, and you were talking about uh a case of uh cyclists being killed by an Uber in 2018 and where the responsibility for that lie, and and that the responsibility, you know, uh effectively lie in the the backup driver and not in the company.
16:51
SPEAKER_00: And and this this is this is gonna be like a big question with chatbots, right?
16:55
SPEAKER_00: I mean, we're you know, we're reading a lot of these like horror stories, you know, certainly like and and there and there will continue to be the more people engage with them, um uh, you know, people like doing like self-harm and things like that.
17:09
SPEAKER_00: So the the the question ultimately is like how and and and can we hold these companies accountable when things go bad?
17:20
SPEAKER_01: Yeah, like the interesting thing to me for all of the responsibility stuff is less like what existing rules do we have?
17:29
SPEAKER_01: Because we have existing rules that assign responsibility in one way or another.
17:34
SPEAKER_01: But it's like the biases that people have in coming up with new rules.
17:40
SPEAKER_01: And um, so for example, that you mentioned the Uber case.
17:44
SPEAKER_01: So like I used to teach this technology ethics class, and we would always discuss this case because it's really interesting.
17:50
SPEAKER_01: There's like a lot of detail, like the company did some stuff wrong, the safety driver did some stuff wrong, the pedestrian did some stuff wrong, like the government could have done more.
17:58
SPEAKER_01: Like you you have basically four, you know entities at fault for what happened.
18:06
SPEAKER_01: And if you ask people who if they had to choose one one entity to blame, who would it be?
18:14
SPEAKER_01: Um, you know, people are pretty split on whether it's the safety driver or the company.
18:19
SPEAKER_01: But what ended up happening in the real world is the company was not held accountable, it was the safety driver.
18:25
SPEAKER_01: And so we just like we discuss this case and how you know we have a bias in society towards holding the safety driver accountable and why.
18:36
SPEAKER_01: And it gets into like the psychology of how we interact with robots and project agency onto them and how the companies become kind of obscured behind that.
18:44
SPEAKER_01: So it's a fascinating space, the responsibility space.
18:48
SPEAKER_01: Um, I know that there are fantastic people working on uh working in that space right now.
18:54
SPEAKER_01: Um, I'm no longer part of it, although we do have a project right now on like safety for humanoid robots that is really fascinating, but it's less about assigning responsibility for harm.
19:04
SPEAKER_00: Well, let's let's talk about that because that's something, you know, I might so you're you're your your sensible bosses are our AI, my sensible bosses are A3, you know, we're doing a lot of regulation in and around uh safety, around um we're around standards around uh robotics, around uh dynamically controlled robots, which are like effectively humanoids.
19:27
SPEAKER_00: Um so what what sort of work is is RAI and are you doing around safety with humanoids?
19:35
SPEAKER_01: Let me back up a little and explain what my team at RAI does because it's kind of a unique thing.
19:44
SPEAKER_01: Um, I don't know of any other robotics organization that has this.
19:47
SPEAKER_01: So, RAI, robotics research organization, we have a team called the Robotics Ethics and Society team that is all social scientists.
19:57
SPEAKER_01: And it's it's this incredible like opportunity to embed social scientists in an engineering institute where we're engaging with engineers every day.
20:08
SPEAKER_01: And when I when I came on to build this team, like one of the things that was important to me was that we weren't just like a traditional ethics team.
20:16
SPEAKER_01: Like we do a lot of the work that you would imagine an ethics team doing, like we do internal education and workshops and we like work directly with engineers and things like task selection and stuff.
20:26
SPEAKER_01: Um, but we also like my my opinion is that there's too many opinions about robots out there.
20:34
SPEAKER_00: So you so you wanted to add one more to kind of override the rest?
20:38
SPEAKER_01: I didn't want to add more opinions.
20:40
SPEAKER_01: I wanted to add more evidence.
20:43
SPEAKER_01: So I specifically hired people who are trained in different methods of data collection and are our kind of outward-facing research.
20:52
SPEAKER_01: A mission is to collect data to better inform decision making and the design or integration or regulation or standard setting or or what have you of robotics.
21:02
SPEAKER_01: And so we do we've done a number of different things.
21:06
SPEAKER_01: Um, like from well, what we can I can't talk about all of the projects yet because they're not all published, and even like the humanoid safety one, although I I can give you kind of the high level what we've been trying to do.
21:18
SPEAKER_01: So, yes, people are thinking about safety in the humanoid space.
21:22
SPEAKER_01: You have a dynamically balancing robot, someone hits an e-stop, what happens?
21:25
SPEAKER_01: Like it's just gonna fall over, right?
21:27
SPEAKER_01: So, what do we do?
21:28
SPEAKER_01: And you know, there's that ISO standard uh being developed.
21:30
SPEAKER_01: So there's fantastic people working on how to make the robots safe.
21:33
SPEAKER_00: Just so yeah, just to for those who don't know, but just to get like drill down a little bit.
21:39
SPEAKER_00: So uh dynamically stable means that when you cut power to it, it's just gonna collapse.
21:44
SPEAKER_00: And that's important.
21:45
SPEAKER_00: So, you know, because then it, you know, it will fall onto something and it could therefore hurt a person.
21:50
SPEAKER_01: Yeah, and it's really important in the humanoid space because the whole promise of humanoids is that they can be in spaces that are built for humans and probably still have human.
22:00
SPEAKER_01: In them as well.
22:00
SPEAKER_01: So it's really important to get the safety right for the technology.
22:05
SPEAKER_01: What we are always interested in as the robotics, ethics, and society team, is talking directly to users and like a different kind of expert.
22:14
SPEAKER_01: So not to the robotics experts, but to workers who have experience working in factories, in um warehouses, experience working with different kinds of robots and other machinery, but have like a really good handle on like what would work best for their space.
22:29
SPEAKER_01: And so we can present them with different options for what the robot should do when you hit the e stop.
22:34
SPEAKER_01: And they have like very strong opinions on what would work and what wouldn't.
22:36
SPEAKER_01: So that's the kind of data that we're collecting.
22:39
SPEAKER_01: We have results from that, but I can't talk about them because we haven't published them yet.
22:42
SPEAKER_01: But like stay tuned.
22:44
SPEAKER_01: Um, that's that's just one kind of project that we do.
22:47
SPEAKER_01: We're really interested in like specific applications and robot integration and understanding what will work best, ideally before the robot has been deployed.
22:58
SPEAKER_01: Although we have also been working with organizations on specific applications where there has been a deployment and looking at how that's changed the nature of what people are doing.
23:07
SPEAKER_01: So we do that type of field work or um experimental work.
23:11
SPEAKER_00: Well, yeah, so this is this is interesting.
23:13
SPEAKER_00: Um maybe I'm putting like way too fine a point on it, but so it's it's data, um, but it it's data uh in the sense that you're going around and collecting data points, but it's the date, the data itself is a collection of opinions, but it's important opinions because it's opinions of people who are working in the field and who are going to work alongside these things.
23:37
SPEAKER_00: And these are things that the people, and then this is like this is this is the conversation we can't keep having over and over again.
23:43
SPEAKER_00: This is a conversation that like just about every single robotics startup has to learn over and over again.
23:49
SPEAKER_00: They all start as roboticists, then like they, oh, this is maybe a good application for ag tech.
23:55
SPEAKER_00: We build a solution, doesn't really make sense, then we have to like learn about agriculture.
24:00
SPEAKER_00: So this is like, hey, what if we started from a point of getting the worker's point of view and then moved out from there?
24:08
SPEAKER_01: Yeah.
24:08
SPEAKER_01: So one of the things that I love about being at Ray is, you know, like you said, like engineers usually start from a place of we have this cool technology, what could be an application for it?
24:20
SPEAKER_01: And as social scientists, we do it the other way around.
24:22
SPEAKER_01: Like you said, we start from the application.
24:24
SPEAKER_01: And so finding the middle of that Venn diagram, that sweet spot where we're making use of like cool new technology, but finding like a really good application for it, that's that's like what excites me most about working with engineers.
24:38
SPEAKER_01: And I think it's like the most uh fruitful avenue to explore for like any robotics company or robotics research institute.
24:48
SPEAKER_01: Um one other thing that we've done in that realm, this is published, so I can talk in more detail about it.
24:56
SPEAKER_01: There's different ways to think about what tasks we might want to automate as a society.
25:01
SPEAKER_01: Like you could think about economic impact, you could think about sustainability.
25:06
SPEAKER_01: Um, one way that roboticists really like to talk about it is dull, dirty, dangerous, the three D's.
25:12
SPEAKER_01: And we looked at like a whole corpus of robotics papers.
25:17
SPEAKER_01: These are mostly research papers, but we we actually also have a project on patents right now where we're seeing a lot of the same.
25:23
SPEAKER_01: Like roboticists love to talk about the three Ds, dull, dirty, dangerous.
25:27
SPEAKER_01: These are the tasks that are undesirable for people to do.
25:30
SPEAKER_01: You know, let's automate those.
25:33
SPEAKER_00: We're not replacing human jobs, or at least the human jobs that we're replacing are the ones that humans don't necessarily want to do.
25:38
SPEAKER_01: Yes, the undesirable ones, right?
25:40
SPEAKER_01: So great, that sounds great.
25:42
SPEAKER_01: Let's start there.
25:43
SPEAKER_01: No one ever actually defines what those tasks or jobs are.
25:47
SPEAKER_01: There seems to be this assumption that we kind of intuitively know which tasks those are.
25:52
SPEAKER_01: But if you look at social science literature and research on the nature of work, you realize that it's actually sometimes not that straightforward and that there's actually like maybe even hidden opportunities for robotics to be helpful.
26:04
SPEAKER_01: But just assuming that we know what those tasks are.
26:06
SPEAKER_01: So, for example, what a dirty task is, that can vary across cultures because dirty is not just about physical dirtiness, it's also about social stigma.
26:16
SPEAKER_01: So if you're a nurse, in some countries, that's like a super honorable profession.
26:21
SPEAKER_01: In other countries, that's dirty work.
26:23
SPEAKER_01: So it's but the cool thing is there's so much data that we have.
26:29
SPEAKER_01: Like we have all of the data on like what tasks are dangerous, um, what workers perceive as dull or repetitive, the prestige of different jobs, like there's so many existing data sources that we can draw on.
26:42
SPEAKER_01: And even though there's some like caveats to that, so like looking at dangerous, for example, you have so many numbers on like injury rates and risk factors of different jobs, like across the world.
26:54
SPEAKER_01: We know this.
26:55
SPEAKER_01: But there are areas where there's like clear underreporting of injuries.
27:00
SPEAKER_01: There are areas where the data is not disaggregated by things like gender, where you might have populations that are like more affected than others, which is like just hidden opportunity for robotics to be helpful, right?
27:11
SPEAKER_01: So, what we did is we come up with better definitions for what dull, dirty, and dangerous tasks are.
27:18
SPEAKER_01: And then we have this framework and like there's even a worksheet that you can go through, like as a company or roboticist, if you care about automating undesirable work or I don't know, motivating your work or marketing it or whatever, you can go through and you can collect data.
27:33
SPEAKER_01: And a lot of the data that we encourage people to uh look for is the worker perspective.
27:40
SPEAKER_01: Um you take an industry like garbage or the recycling industry, and like intuitively we think, oh, yeah, that's dull, dirty, and dangerous.
27:49
SPEAKER_01: Maybe we should automate like trash pickup collection.
27:52
SPEAKER_01: But then you look and there's all these studies of like sanitation workers in different cities, and some of the proposals for automating their work and making it safer actually also take away the one aspect of the job that they find enjoyable.
28:05
SPEAKER_01: And if you know that, you can design for that, right?
28:08
SPEAKER_01: How can we make it safer, but also not take away the one thing that they enjoy about the job?
28:13
SPEAKER_01: So, anyway, I I'm really passionate about that project.
28:16
SPEAKER_00: It's just a little one, but no, that's really that's really interesting because I I hadn't really contextualized it that way.
28:22
SPEAKER_00: And and and as you're speaking about it, it occurs to me that the way that we define old, dirty, and dangerous are the jobs that roboticists don't want to do.
28:31
SPEAKER_00: Effectively.
28:33
SPEAKER_01: It's funny, like mean people have said things like because robotics is full of you know, young male engineers who have disposable income, and mean people have said that they're just trying to automate, you know, the tasks that their mom used to do, like the chores that they don't want to do, like the laundry folding and stuff.
28:50
SPEAKER_01: But you know, I I do no, I see a lot of people who care a lot about they want to have that impact, but like we all have blind spots, right?
28:59
SPEAKER_00: Yeah, no, that I mean that that and and that's a fair, that's a fair point, also like across the board as far as automation.
29:06
SPEAKER_00: Because this is this is the conversation that we're having too, is like that automation doesn't replace jobs, right?
29:11
SPEAKER_00: That people are saying that like that that the jaws shift.
29:13
SPEAKER_00: But the question is ultimately, is automation going to make them these jobs better, like the promise is, or in the case of like dishwashers and other automation that we've seen in the past, like does it actually make it the do the standards raise?
29:26
SPEAKER_00: And like, can automation ultimately we've seen this in the warehouse, right?
29:29
SPEAKER_00: Like we've seen the precision with which we can track people like picking and placing objects, like makes the job that much more difficult.
29:37
SPEAKER_00: So, like ultimately, is automation definitely making jobs easier, or can it make it more like, you know, can it ultimately make it more difficult for humans?
29:48
SPEAKER_01: I think one of the main mistakes we make is just viewing this in this deterministic way, like what's going to happen when really we have so much choice?
29:58
SPEAKER_01: And by we, I mean we as a society, because I think that, you know, this narrative, the robots are gonna take the jobs, like there's still like headlines like that everywhere.
30:09
SPEAKER_01: That's not really accurate in describing, you know, robots, robots aren't coming for the jobs.
30:14
SPEAKER_01: What's what's coming for the jobs is company decision making against the backdrop of unbridled, you know, capitalism and globalization and a political economy that treats workers like replaceable commodities.
30:25
SPEAKER_01: Like that's that's the problem.
30:27
SPEAKER_01: Um, that's a difficult thing to address and change, but that's that's what we need to be talking about.
30:32
SPEAKER_01: And I think that, but even against this backdrop, like in our work at Ray, I can't name any companies, but we work directly with companies on our team.
30:44
SPEAKER_01: And we see that there's a lot of choice that people have in terms of thinking about how to automate things.
30:52
SPEAKER_01: You can say, oh, we want to take this robot and we want the robot to do exactly what the person is doing, and then we can lay off the person.
30:59
SPEAKER_01: Or you can be thinking more creatively about like your entire workflow and how you can introduce automation to help make it more efficient and increase quality control, for example.
31:09
SPEAKER_01: So there are so many, it takes more creativity, it takes more longer-term thinking rather than just thinking shorter term about saving a quick buck by replacing a person.
31:20
SPEAKER_01: But I think the companies that are putting the effort into that thinking are ultimately going to get more out of it.
31:26
SPEAKER_01: This isn't just about, you know, you know, let's be nice to the workers.
31:31
SPEAKER_01: This is also about how do we think creatively about a future where robotics is actually helpful to the whole, you know, thing we're trying to achieve rather than just replacing something we already have.
31:43
SPEAKER_00: Yeah, and this actually somewhat brings us back to the Uber conversation in an interesting way, because that was being discussed in the context of regulation.
31:53
SPEAKER_00: And the point, one of the points that you were making in that conversation was that certainly in in the US, in the context of like free market thinking, regulation tends to be thought of as being like anti-innovation, as being like stifling of innovation.
32:09
SPEAKER_00: And your point was that that doesn't necessarily have to be the case and often like isn't the case.
32:14
SPEAKER_00: If you point, I think you pointed to the EU as maybe like a good example of you know regulation.
32:20
SPEAKER_00: No, okay.
32:22
SPEAKER_00: Maybe not.
32:23
SPEAKER_00: No, no, but point being that that that uh that that regulation like doesn't necessarily like need to hamper or stifle innovation.
32:33
SPEAKER_01: No, like in some cases, companies actually want regulation and because they want to know what to design for, because they know that something's coming down the line.
32:45
SPEAKER_01: So um, but and and also just like I I don't, yeah, I don't I don't think it's like anti-innovation to to have regulation.
32:55
SPEAKER_01: I think regulation when done right, and that's a big if, right?
33:00
SPEAKER_01: Because there's a ton that goes wrong.
33:03
SPEAKER_01: As someone who has a legal background, like there's there's a ton that we could be doing a lot better.
33:08
SPEAKER_01: Um, and regulation is also like way behind the times and like behind the you know current state of the art and technology and all of that.
33:15
SPEAKER_01: Like all of that is a problem, yes.
33:17
SPEAKER_01: But in theory, you could regulate in a way that actually drives innovation in a direction that is societally valuable and is not stifling innovation.
33:27
SPEAKER_01: Like you can nudge things in a good direction.
33:30
SPEAKER_01: So I I'm still a believer in regulation, even though we need to improve how we're doing it.
33:36
SPEAKER_00: This is interesting.
33:37
SPEAKER_00: So um, I I don't know if you know like anything about my background, but before I started really writing about robotics and automation, I was writing a lot about uh consumer technology.
33:45
SPEAKER_00: Um and uh one one of the things it write about like phones and and laptops, and what one of the things we would do is comb through all the patents.
33:55
SPEAKER_00: And it was very clear that the people who are judging these patents like didn't really know a lot about technology and like understandable, right?
34:04
SPEAKER_00: I mean, if you know you're a judge who's been sitting on a bench for a really long time, like you probably don't necessarily understand the nuances of technology.
34:12
SPEAKER_00: And that's probably also the case for a lot of these regulations that we're talking about right now.
34:18
SPEAKER_00: Do you do you see that as potentially being part of like the role that you can play is sort of helping kind of address some of these, or maybe like getting out ahead of the eight ball when it comes to some of these maybe future concerns?
34:31
SPEAKER_01: Yeah, so right now our team is not working on policy as much.
34:36
SPEAKER_01: We're really focused on working with um industry partners and more like local communities, underserved communities.
34:43
SPEAKER_01: But we did have one project where there was an opportunity to like actually talk to local government and make them aware of the state of the art of the technology as they were trying to create like permits and standards.
34:59
SPEAKER_01: So, and I think it's incredibly important.
35:01
SPEAKER_01: I think that unfortunately, it's like rare to have people who understand policymaking and are also like really at the cutting edge of technology and can communicate that.
35:13
SPEAKER_01: So we need more.
35:15
SPEAKER_01: It's I will say it's gotten a lot better.
35:17
SPEAKER_01: Like it was like 14 or 15 years ago that you know, some people started this lawn robotics conference that I was a part of.
35:26
SPEAKER_01: And at the time, people in the legal field said that we were just doing sci-fi stuff and we were just like having a conference to like talk about cool robots.
35:35
SPEAKER_01: Why would we need legislation for robots and AI?
35:38
SPEAKER_00: Like to be fair, having a conference to talk about cool robots is like a totally admirable goal, too.
35:42
SPEAKER_01: I mean, it can be both, right?
35:44
SPEAKER_01: Like, but like we also believe that it was important, and now like everyone sees that it is important, and I think that there's a lot more of that interdisciplinarity where you have technologists talking to policymakers.
35:58
SPEAKER_01: I think like we're not anywhere close to where we need to be, but at least it's changing.
36:03
SPEAKER_01: And I do like I do see that as like a lot of my role.
36:08
SPEAKER_01: I want to at least try to make technology understandable to anyone, um, but also to policymakers.
36:16
SPEAKER_01: So I think that's so important.
36:19
SPEAKER_00: Somebody I just spoke with actually for as as we're recording this, it it just went out, I think, that in this week's newsletter is Austonian Karen Panetta, who's been doing some very uh interesting work.
36:30
SPEAKER_00: Are you familiar with her?
36:30
SPEAKER_00: She's at Tufts University uh at IEEE.
36:33
SPEAKER_00: She's been doing uh interesting work work in and around sort of like like ethics and and AI for a long time now.
36:40
SPEAKER_00: Uh I should introduce you.
36:41
SPEAKER_00: I feel like the two of you would have a lot to talk about.
36:44
SPEAKER_00: But one of one of the things, one of the things that she she talked about that that I considered, because we were going to have a conversation about uh biases and physical AI, right?
36:56
SPEAKER_00: And and going into the conversation, I was sort of thinking about it.
37:01
SPEAKER_00: Obviously, like the SIL factors into it, but the the similar to the conversations that we had have have been having around biases and just general AI, you know, um, as far as like uh like cultural biases or or gender biases.
37:14
SPEAKER_00: But she she brought up the point that you know when we're talking about physical AI, like accessibility has to be a really big part of it as well, right?
37:22
SPEAKER_00: Because all of a sudden we're entering the physical world.
37:25
SPEAKER_00: When when you're uh talking about these uh underrepresented communities that you're working with, I mean, um who who is who is being affected and and and who are you sort of, I guess, advocating for?
37:40
SPEAKER_01: Yeah, so we have trying to think of how much I could say.
37:45
SPEAKER_01: Um we've worked with multiple communities of people who are going to encounter technology or have to use technol robot technology without being involved in like the design of it.
38:00
SPEAKER_01: So and I think one thing that's really important to our team is understanding that like when you deploy a robot into you know a situation, like you really have to understand the whole context in order to make sure that the robot is actually helpful.
38:18
SPEAKER_01: Like we're talking about a socio-technical system, we're not talking about like you just put the robot in and it works or it doesn't.
38:24
SPEAKER_01: Like there's a whole context of human behavior and norms and organizational decision making and culture that you are putting that robot into.
38:31
SPEAKER_01: And if you don't understand that, then it can go really wrong.
38:35
SPEAKER_01: So we have looked at like certain applications that have been proposed for like helping certain communities with robots.
38:43
SPEAKER_01: And we've gone in and done like extensive interviews with people, like stakeholders in that community, like trying to understand how well the systems currently work, what could be the pain points for all of this, so that we get that information before you know the robot is deployed and fails, right?
39:00
SPEAKER_01: Um, so that's that's some of the work that we've been doing.
39:03
SPEAKER_01: Um it's been great.
39:05
SPEAKER_01: Right now we're more focused on uh workers, so factories, manufacturing contexts, warehouses, um, because that's also another like user group that is extremely affected by robotic technology, but isn't always consulted.
39:21
SPEAKER_01: Like the people will deploy technology and they'll consult like at the management level, um, or at the like factory manager level, but they won't talk to the workers on the floor.
39:33
SPEAKER_01: In part sometimes there's a language barrier.
39:35
SPEAKER_01: Sometimes they just don't like consider, they they consider the factory manager like as someone who just knows it all.
39:40
SPEAKER_01: But we're finding that there's a lot of very rich information you can get from the people on the floor um that is actually helpful in like uh choosing how to design or use the robot in practice.
39:53
SPEAKER_00: Yeah, I think this is this is really interesting because you know, right now I keep having this conversation around scaling and all the difficulties, like specifically around humanoids.
40:04
SPEAKER_00: Um, you know, and when we talk about scaling, you know, we're talking about um, you know, all the manufacturing concerns and um, you know, actually like uh deploying these out in the world.
40:18
SPEAKER_00: But then there are a lot of like human concerns that we have and that like that that are also gonna ultimately be a part of the conversation that we're not really speaking about right now.
40:28
SPEAKER_00: And they're similar.
40:29
SPEAKER_00: It strikes me that they're similar to these like lessons that technologists keep learning over and over again.
40:35
SPEAKER_00: You know, we were we recently spoke to the VP of Alexa, right?
40:40
SPEAKER_00: And it's like, well, yeah, I mean, you've got a smart assistant and you're rolling it out all over the world.
40:45
SPEAKER_00: Like you need people locally in different cultures to understand those cultures because there are a lot of nuances that you have to understand in order to roll them out.
40:56
SPEAKER_00: So like somebody, somebody is going to have to play that role, whether it's RAI or someone else, in order to like to bridge that gap and make sure that human beings are being consulted in the process of automating.
41:11
SPEAKER_01: Yeah, and like social scientists are trained to do that, right?
41:14
SPEAKER_01: Yeah, and like our our team, like it was also important to me to not like just hire Americans too, because like America's not the center of the universe and well, agree to disagree.
41:26
SPEAKER_01: Oh, really?
41:27
SPEAKER_01: No, the center of your universe, but like I don't know, it just seems yeah, like I, you know, you know, some of my best friends are roboticists, right?
41:37
SPEAKER_01: And like I'll watch them walk walk into a kitchen and pick up a cleaning rag and they're looking at it, and I'm like, I can see exactly what they're thinking.
41:45
SPEAKER_01: They're like, this would be hard for a robot to do.
41:47
SPEAKER_01: Maybe I can train the robot to do this.
41:49
SPEAKER_01: It's like that's great, but you know, maybe there's some application that you're not even thinking of because if you consulted with people in other countries, or like, oh, recently we did, I think recently we did a workshop with um teenagers from like um like a underrepresented community like um after school program.
42:11
SPEAKER_01: And we asked them to just like envision, like with no technical technical constraints, envision applications of robots that would be like cool for for them to have in their lives.
42:23
SPEAKER_01: And it was just so different from some of the applications that you know my roboticist friends are envisioning.
42:30
SPEAKER_01: So I don't know, I think there's so much to be said for Do you have any examples that you can share?
42:35
SPEAKER_01: Yeah, oh I guess, well, so we have we didn't do a direct scientific comparison of anything.
42:42
SPEAKER_01: Um, but you know, there's artwork all over Institute and on our homepage showing certain like possible future applications for robots.
42:51
SPEAKER_01: And it's like robots fixing your car and robots fixing your, you know, laundry machine, or robots doing this or that.
42:58
SPEAKER_01: A lot of what these kids came up with was more community focused, like um cleaning um like community spaces or like environmental stuff or involving parks and libraries.
43:12
SPEAKER_01: So it was just like a very different um kind of cultural, like do you know what I mean?
43:20
SPEAKER_00: Yeah, it's literally, it's literally stuff that it's more difficult to monetize.
43:25
SPEAKER_01: Yes, yes, but also like Ray isn't about monetizing anything.
43:31
SPEAKER_00: But but my my my point, no, my my my point is like like that's a big part of the reason why that's so top of mind for all these roboticists is because they have to start thinking of like how to get that return investment up front.
43:45
SPEAKER_01: Maybe.
43:47
SPEAKER_01: I mean, we have stuff in our artwork that would be impossible to monetize, if you ask me.
43:53
SPEAKER_00: That's fair.
43:54
SPEAKER_00: Like, like But it's it's it is it is ingrained in the culture, like certainly like startup culture and everything else, that like, all right, this.
44:00
SPEAKER_00: Needs to like start generating money.
44:02
unknown: Okay.
44:02
SPEAKER_01: And now I just feel like I need to defend roboticists because like you did not expect that turn in this conversation.
44:10
SPEAKER_01: Yeah.
44:11
SPEAKER_01: It's not that like money isn't like one, you know, sure way to frame the value of something, but like I see, especially younger generations of engineers, like especially coming straight out of school, coming into a place like this, and they're like, I actually want to help people.
44:27
SPEAKER_01: Like I want my the technology I work on to be helpful, not just make money.
44:33
SPEAKER_01: Um and they're looking for guidance on how to do that because they're just trained to make the technology.
44:41
SPEAKER_01: And I think one of the like amazing things about being here has also been that I think social scientists are often, or like especially, especially in like the tech criticism space, we're always like, oh, you know, roboticists need to be more responsible and thinking about like how to make their technology actually helpful, blah, blah, blah.
44:59
SPEAKER_01: I feel like engineers have really absorbed that message.
45:01
SPEAKER_01: Like they come in here wanting to do good, and they're like, okay, so tell us how to do it.
45:07
SPEAKER_01: And then the social scientists are like, uh, well, because no one understands, like I think we do, like our team does, because we've been working with the engineers for two and a half years.
45:18
SPEAKER_01: We understand their workflows, we understand what like technologies they're considering and what the trade-offs are.
45:24
SPEAKER_01: But for most people in our field, like they have no idea how to be constructive.
45:27
SPEAKER_01: They're just like, don't do that, you know?
45:29
SPEAKER_01: So I don't know again, like it's so important to have that interaction of the the disciplines.
45:35
SPEAKER_01: We really need more.
45:36
SPEAKER_00: Yeah, that that this is a really interesting topic of conversation for me.
45:41
SPEAKER_00: I was a, I don't know how I ended up here.
45:43
SPEAKER_00: I was a creative writing major at UC Santa Cruz.
45:47
SPEAKER_00: Like I it was a really weird path to get here.
45:50
SPEAKER_00: Um, and I'm always fascinated when I'm looking at somebody's CV and they ended up being uh, you know, a roboticist or the CEO of robotics company, and you know, they have a history and and philosophy or or some other uh humanity.
46:07
SPEAKER_00: And and and I I'm I'm I'm curious to get your take.
46:11
SPEAKER_00: Uh you know, as you're saying, like this this overlap between uh technology and social sciences is really important, but uh how important it is that uh people be exposed, like the the technologically minded people, the people who are building these systems are directly exposed to those uh ideas uh during their I guess development process.
46:37
SPEAKER_01: I mean, I think it's important I'm biased, right?
46:40
SPEAKER_01: Sure, but I think it's important and like Brian, like being at Ray, like we will have conversations with people like by the coffee machine or at lunch, where it's just like such low-hanging fruit.
46:52
SPEAKER_01: Like we're both coming from such different worlds and we ask different questions.
46:56
SPEAKER_01: Like, so you know, we'll be like, hey, have you considered this?
47:00
SPEAKER_01: And the roboticists will be like, oh yeah, but that doesn't work because of this.
47:03
SPEAKER_01: And like there's this back and forth that like ends up helping everyone.
47:08
SPEAKER_01: Like, we need more of this.
47:10
SPEAKER_01: Um, and yes, like people should be exposed to it.
47:13
SPEAKER_01: Ideally, you would have people working together on it.
47:16
SPEAKER_01: That's my dream.
47:17
SPEAKER_01: Like to have social scientists working with engineers on these projects instead of living in two separate worlds.
47:23
SPEAKER_00: Yeah, I mean, really backing the camera out, because I this is something I we didn't really touch on, which we should.
47:29
SPEAKER_00: Um abstractly, I think probably most people listening to us kind of understand what RAI is and and and where they fit in, you know, as compared to like maybe like a TRI, you know, where it's it's funded by by Hyundai, it's not uh a commercial outfit.
47:45
SPEAKER_00: Um, they're they're doing all of this research, but I guess broadly speaking, what do you see the role of the institute being?
47:55
SPEAKER_01: Oh, I think it's so important to have this type of research organization.
47:59
SPEAKER_01: Like it's having a fully funded research organization that is able to do longer-term things and more robust things than a university, for example.
48:11
SPEAKER_01: Like I, you know, I was at MIT, MIT's awesome, right?
48:14
SPEAKER_01: But a lot of the robotics projects are just like duct taped together.
48:17
SPEAKER_01: They're like a proof of concept and like it it being at a research organization that has like a cross-functional hardware team that is building real robots and making them like actually robust and and you can replicate them as stuff is like super cool and not something that a lot of people in academia are doing.
48:34
SPEAKER_01: And then on the other hand, industry is like very product focused, right?
48:37
SPEAKER_01: And there's short time horizons and you got to get the product out.
48:40
SPEAKER_01: So they don't really have the space or the means to do the forward-thinking research or the more like at a level of abstraction, like how can we improve the ability of robots to do X like broadly instead of just focused on like a specific, you know, assembly line task.
49:00
SPEAKER_01: Um, so it's just like there's a gap.
49:02
SPEAKER_01: I think the challenge is always getting, you know, getting these types of institutions funding, but you know, it's work before.
49:09
SPEAKER_01: Like I think Mark, uh, when he started Ray, called it the Bell Labs of AI and robotics.
49:15
SPEAKER_01: And I think it's really necessary to have that type of organization.
49:18
SPEAKER_00: Yeah, I I think maybe something like a lot of people, the people who didn't um study at the schools, like don't maybe uh understand or appreciate.
49:26
SPEAKER_00: And this is something I didn't appreciate until I started visiting them was yeah, like Ivy League schools, really high tuitions.
49:34
SPEAKER_00: We tend to like as a humanities major, I tend to think of the sciences and technologies as being really well funded.
49:40
SPEAKER_00: And then you go into the labs, and yeah, everything is like it's like a connected duct taped to the back of like a now robot or something like that.
49:47
SPEAKER_00: It's like shocking, like how how McGyvered everything is together.
49:52
SPEAKER_01: Yeah, like I I remember the first time I visited Boston Dynamics.
49:57
SPEAKER_01: Like I came in and it was it was Mark actually who was showing me around.
50:03
SPEAKER_01: And um, there was an Atlas robot just kind of like slumped against the wall somewhere.
50:08
SPEAKER_01: And I tried to make a joke.
50:09
SPEAKER_01: I was like, oh, it's just like an MIT, the robots aren't actually working.
50:13
SPEAKER_01: And then he just like wordlessly opened the door into the next room, and they had probably like 20 spot robots just walking around this like obstacle course fully autonomously, and with no people there, and my jaw dropped.
50:28
SPEAKER_01: I was like, it was like the coolest thing I'd ever seen.
50:32
SPEAKER_01: Um it's just yeah, universities don't like there's no point in you're doing research, you have grad students coming in and out.
50:41
SPEAKER_01: There's no point in developing like a sustainable platform.
50:45
SPEAKER_00: It makes me feel better that you make the same first impression on people that I think you make, which is just like a totally misfired joke.
50:53
SPEAKER_00: But then just sticking my foot in my mouth.
50:57
SPEAKER_00: Yeah, yeah, yeah.
50:58
SPEAKER_00: Um we're we're we're really close to time, and and we I promised that we would get to Plio at the top of the conversation, and we haven't really, and I really want to talk about Plio.
51:08
SPEAKER_00: One because like Plio was was such a cool robot.
51:12
SPEAKER_00: Um we're we're gonna have to have the the creator on at some point, but Plio, there there are there are a lot of pictures of you with Plio and Plio in various states of undress, like skeletal Plio.
51:23
SPEAKER_00: Uh, and then Plio plays a very pivotal role in in your new breed book.
51:28
SPEAKER_00: And in fact, there is one scene in the book that has really stuck with me, in which it's you're teaching a workshop and you it's effectively like a you force someone to ritualistically sacrifice a Plio robot.
51:45
SPEAKER_00: Can you talk a little bit about what led up to that moment?
51:51
SPEAKER_01: Sure.
51:52
SPEAKER_01: And like in my defense, like we didn't know.
51:55
SPEAKER_01: We didn't know it was gonna be that dramatic.
51:57
SPEAKER_01: Um, yeah, so PLO PLIO is the whole reason that I got interested in robotics.
52:06
SPEAKER_01: Um I bought a PLO just like out of I'd always loved robots, but I was still like a law student at the time, and I bought a PLO, it came out what in 2007, I think.
52:16
SPEAKER_01: And it was also like you said, it's such a cool robot, has all these motors and touch sensors, and then like you're gonna be able to do it.
52:22
SPEAKER_00: It's a dinosaur, it's a living dinosaur.
52:25
SPEAKER_00: What's not like dinosaur?
52:26
SPEAKER_01: It looks like the land before time, and it's like a baby chimerosaurus, and it mimics life like behavior really well and emotions and stuff, and so it has a tilt sensor.
52:35
SPEAKER_01: So if you hold it up by the tail, it'll start crying, um, and and like squirming.
52:40
SPEAKER_01: And I thought this was really interesting and was showing it to my friends, and like they would one person held it up for a really long time and just like watched it squirm, and it made me really uncomfortable.
52:52
SPEAKER_01: And I was like, okay, put the robot back down now, that's enough.
52:56
SPEAKER_01: So that was kind of like that was the first moment where I was like, oh, that's interesting.
53:01
SPEAKER_01: Like, I know exactly how this thing works, but like I feel really compelled to be kind to it.
53:06
SPEAKER_01: Is this just me?
53:06
SPEAKER_01: What's going on?
53:08
SPEAKER_01: And then fast forward a few years, we did this workshop that you mentioned, this now infamous workshop where we gave people five of these baby dinosaur robots and had them name them and play with them and interact with them.
53:20
SPEAKER_01: And then we unveiled a hammer and a hatchet and a knife, and we were like, okay, like we're gonna start like quote unquote hurting the robots.
53:28
SPEAKER_01: And they just like totally refused to do it.
53:31
SPEAKER_01: Like we thought, initially, we thought, okay, there's probably gonna be a split in the room between people who are like, sure, like it's just a robot and can't pull anything.
53:39
SPEAKER_01: Yeah.
53:39
SPEAKER_01: Yeah.
53:40
SPEAKER_01: Right.
53:40
SPEAKER_01: Like, and and people who are have more of an issue with what we were asking them to do, but they all just like totally refused.
53:46
SPEAKER_01: They were like, no.
53:47
SPEAKER_01: Um, so we improvised.
53:49
SPEAKER_01: Uh, this this was uh, I was doing this with my colleague Hannes Gossel, and I think it was actually his idea.
53:55
SPEAKER_01: He was like, okay, let's um let's tell them they can save their team's robot if they hit another team's robot with a hammer.
54:03
SPEAKER_01: And so we tried to do that with them for a while and they totally couldn't do it.
54:06
SPEAKER_01: Um and finally we threatened to destroy all the robots unless someone took a hatchet to one of them, and some guy did.
54:14
SPEAKER_01: And, you know, it was it was just so dramatic.
54:17
SPEAKER_01: Everyone like stood in a circle and kind of winced and looked away as he brought the hatchet down on the, you know, cute baby dinosaur robot's neck.
54:24
SPEAKER_01: And then there was this like half-joking, half serious moment of silence for the fallen robot.
54:30
SPEAKER_01: And so, like, this was a workshop.
54:33
SPEAKER_01: It wasn't a scientific setting.
54:36
SPEAKER_01: There were no conclusions we could draw from it.
54:38
SPEAKER_01: There were social dynamics, people really got into this narrative and stuff.
54:41
SPEAKER_01: Um, but it was kind of a career-changing moment for me because at that point I was like, okay, there's something really interesting here.
54:48
SPEAKER_01: And so I pivoted from what I was doing and started doing human-robot interaction research at MIT.
54:55
SPEAKER_01: And I started by doing some work with Cynthia Brazil, who's the creator of Jibo, and Palash Nandi, who's now at Google.
55:04
SPEAKER_01: And we were looking at different factors that could influence people's like hesitation to want to quote unquote harm like a very lifelike behaving robot.
55:13
SPEAKER_01: And we were looking at things like people's tendencies for empathy and all of that.
55:16
SPEAKER_01: So that like set me down the path of the social emotional connection that people have with robots and why we treat robots like they're alive, even though we know that they're just machines.
55:25
SPEAKER_01: And I still have I have so many Pleos at home.
55:28
SPEAKER_01: Like the skeleton one you mentioned is actually an early prototype that someone from the someone who had previously been at the company sent me.
55:35
SPEAKER_01: I actually finally met Caleb Chung, the creator, um, a few months ago.
55:39
SPEAKER_01: Like, and it was like amazing.
55:42
SPEAKER_01: You should totally have him on.
55:43
SPEAKER_01: He's brilliant.
55:44
SPEAKER_01: I love him.
55:45
SPEAKER_00: Yeah.
55:46
SPEAKER_00: So maybe I don't know.
55:47
SPEAKER_00: Maybe um maybe this is the next book.
55:49
SPEAKER_00: I mean, like what the what what are the implications?
55:52
SPEAKER_00: Like, once we do get to a point as a society, whether it's the familiar machines or magic, or whoever like ultimately cracks that code where these robots do become pets, you know, what what what does it mean to be able to compartmentalize to, you know, to be able to kill it with a hammer, but also still have it as a pet.
56:13
SPEAKER_01: You know what's really fascinating to me as like a next step.
56:17
SPEAKER_01: So there's so much research by myself, but many other people on like how we treat these robots like living things, but there's not enough research on like what's actually different and special about them.
56:30
SPEAKER_01: Like we know that, for example, people interact with chatbots and robots, like they will, you know, anthropomorphize them and like treat them like a living thing, but also they understand that they're interacting with the machine.
56:42
SPEAKER_01: So, for example, they feel less judged and they will share more information.
56:47
SPEAKER_01: Um, so there's there's all these differences in how we interact with robots as well.
56:51
SPEAKER_01: And I think if social robotists could lean into those differences and what really makes the robots uniquely valuable, that would be interesting.
57:00
SPEAKER_00: Well, we're at time.
57:01
SPEAKER_00: Uh, I'm glad we finally got a chance to do this, Kate.
57:04
SPEAKER_00: Thank you so much.
57:05
SPEAKER_01: Thank you.
57:06
SPEAKER_01: This is awesome.
57:08
SPEAKER_00: Thank you to Kate and RAI.
57:10
SPEAKER_00: Thanks you as ever for tuning in and making it deep into this program.
57:15
SPEAKER_00: If you enjoyed that experience, please like and subscribe and check out more content over at automated.fm.
57:22
SPEAKER_00: And we will see you next week for another episode of Automated.