WEBVTT - Robots Getting Social 0:00:00.160 --> 0:00:07.080 Brought to you by Toyota. Let's go places. Welcome to 0:00:07.280 --> 0:00:14.520 Forward Thinking, pay there, and Welcome Beforeward Thinking, the podcast 0:00:14.600 --> 0:00:16.759 that looks at the future and says people are strange 0:00:16.800 --> 0:00:20.280 when you're a stranger. I'm Jonathan Strickland, I'm La, and 0:00:20.400 --> 0:00:23.200 I'm Joe McCormick, and I've got a story for y'all. Alright, 0:00:24.239 --> 0:00:28.480 let's gather around the campfire. It's more of an anecodote. Actually, 0:00:28.520 --> 0:00:31.800 it's not even that anyway, here's what it is. You've 0:00:31.840 --> 0:00:34.080 built it up so much and now I'm already bound 0:00:34.159 --> 0:00:37.800 for disappointment. No please. So a while back I was 0:00:37.800 --> 0:00:42.159 writing a video episode for Forward Thinking about whether robots 0:00:42.159 --> 0:00:44.640 are going to take our jobs? And if you haven't 0:00:44.720 --> 0:00:47.440 doing that video, yeah you should go watch it. It's 0:00:47.440 --> 0:00:50.279 on YouTube. It's a great, great episode. It is. It is. 0:00:50.360 --> 0:00:52.760 But we have a spoiler for that video. Yeah, there 0:00:52.800 --> 0:00:54.840 is a spoiler. Actually this We say this in the 0:00:54.920 --> 0:00:58.160 video pretty early on. It's not the end. Yes, they will. 0:00:58.480 --> 0:01:00.960 Robots will take our jobs, will take your job, They'll 0:01:01.000 --> 0:01:03.800 take our jobs. They'll they'll take all the jobs. But 0:01:04.400 --> 0:01:07.360 in doing the research for the episode, I came across 0:01:07.520 --> 0:01:10.640 plenty of good reasons for thinking that some jobs are 0:01:10.959 --> 0:01:15.720 much much harder to automate than others. And one simple 0:01:15.760 --> 0:01:19.720 rubric for separating jobs into robot friendly and robot unfriendly 0:01:20.000 --> 0:01:24.440 is this is the job easily described in an explicit 0:01:24.600 --> 0:01:28.160 list of instructions that can be executed over and over. So, 0:01:28.200 --> 0:01:30.760 in other words, could it be something like pick up 0:01:30.800 --> 0:01:36.000 blue box, open blue box, put orange sphere in blue box, 0:01:36.120 --> 0:01:40.119 close blue box. That's easily explained, and it would be 0:01:40.200 --> 0:01:43.240 something that you could program as a list of actions 0:01:43.640 --> 0:01:48.640 for a robot that has very simple image recognition software. 0:01:48.720 --> 0:01:51.080 It really just needs to know what a blue box 0:01:51.120 --> 0:01:53.280 looks like and what an orange sphere looks like, and 0:01:53.280 --> 0:01:55.680 then you're pretty much good. Hard to touch those things 0:01:55.920 --> 0:01:58.800 that they don't shatter whatever, right, and it's it's just 0:01:58.840 --> 0:02:01.640 gonna do the same things over and over again. There 0:02:01.680 --> 0:02:04.480 aren't going to be a lot of changing external conditions, 0:02:04.600 --> 0:02:07.800 or if there are, they might not matter necessarily. Jobs 0:02:07.880 --> 0:02:10.920 like this are pretty easy to automate, and they they 0:02:11.000 --> 0:02:13.920 if if they have not already gone to machines or 0:02:14.080 --> 0:02:17.480 computer programs or robots, they probably will soon. So a 0:02:17.480 --> 0:02:20.360 lot of examples of this might be jobs today, and 0:02:20.400 --> 0:02:24.840 like data entry or data processing or jobs like telemarketing. 0:02:25.440 --> 0:02:27.400 A lot of people think that these kind of jobs, 0:02:27.520 --> 0:02:30.720 because they're so repetitive and you can make clear lists 0:02:30.720 --> 0:02:34.519 of instructions, can pretty easily be done by machines. Other 0:02:34.639 --> 0:02:39.560 jobs are very robot unfriendly. They're harder to automate because 0:02:39.600 --> 0:02:42.640 they require things that are harder to predict, Like you 0:02:42.680 --> 0:02:47.079 can't write a program that does unpredictable things. I mean, 0:02:47.400 --> 0:02:51.400 programs are pretty much by the definition of predicting. And 0:02:51.400 --> 0:02:55.080 we've talked about random number generators and how difficult it 0:02:55.160 --> 0:02:58.639 is to have a truly random number generator. Yeah. Yeah. 0:02:58.639 --> 0:03:01.360 If if you say to a robot, literally any person 0:03:01.840 --> 0:03:04.520 who exists in the world could walk through that door, 0:03:05.840 --> 0:03:08.040 how do you deal with that person going to go 0:03:08.080 --> 0:03:11.520 like and just shut down? Yeah? Yeah, yeah, So these 0:03:11.600 --> 0:03:14.000 jobs are the one The hard to automate jobs are 0:03:14.040 --> 0:03:20.280 the ones that require creativity, strategizing, reacting to unpredictable circumstances 0:03:20.320 --> 0:03:23.280 and stuff like that. And there are some jobs I 0:03:23.320 --> 0:03:26.760 thought it was interesting that sound like they fit into 0:03:26.760 --> 0:03:29.600 the first category, but they actually fit better into the 0:03:29.639 --> 0:03:31.720 second well. And there are also some that sound like 0:03:31.760 --> 0:03:33.760 they would fit in the second but seem to fit 0:03:33.800 --> 0:03:37.200 just fine the first like you mentioned telemarketing. Well, telemarketing 0:03:37.280 --> 0:03:41.480 involves having conversations with actual people, and you might think, oh, 0:03:41.520 --> 0:03:45.360 well a robot would have different would there be some unpredictability, 0:03:45.400 --> 0:03:48.680 But often in those cases there's a very specific script 0:03:49.040 --> 0:03:52.120 to follow, whether you're human or robot, and you have 0:03:52.240 --> 0:03:54.560 a very it's essentially kind of a it's a it's 0:03:54.600 --> 0:03:57.800 a very simple decision tree, right, You're trying to get 0:03:58.080 --> 0:04:01.200 the person on the line to say yes or or 0:04:01.320 --> 0:04:05.680 no in some cases, and if as long as it continues, 0:04:05.720 --> 0:04:07.520 it can keep going down that pathway. And we've even 0:04:07.520 --> 0:04:12.640 seen some interesting clever uses of natural language recognition to 0:04:12.800 --> 0:04:17.080 create situations where a person is not entirely sure if 0:04:17.120 --> 0:04:19.560 the if they're actually speaking to a human being or 0:04:19.600 --> 0:04:22.480 a robot telemarketer. Yeah, and that might be some cases 0:04:22.480 --> 0:04:24.800 where even if it is really a human, they're they're 0:04:24.800 --> 0:04:27.240 like trying to get the human to behave like a robot. 0:04:27.320 --> 0:04:31.320 So these kind of robo calls like that, that's prime 0:04:31.440 --> 0:04:36.080 territory for automation, especially also if it's low stakes, you know, 0:04:36.120 --> 0:04:40.200 because the people the bosses like aren't terrified about what's 0:04:40.240 --> 0:04:42.640 going to happen if the person being called has a 0:04:42.640 --> 0:04:48.000 bad experience. You know, it's just that I don't care, right, right, 0:04:48.360 --> 0:04:51.760 But let's take an example of something from our favorite movie, 0:04:52.200 --> 0:04:56.320 Back to the Future Part two. Right, So, in the 0:04:56.320 --> 0:04:58.919 opposite category is stuff that seems like it might be 0:04:58.960 --> 0:05:01.120 easy to automate, but that actually it would probably be 0:05:01.240 --> 0:05:06.080 really hard to automate, And like robot servers in a restaurant. Yeah, 0:05:06.160 --> 0:05:08.680 so a robot server in a restaurant, what would it 0:05:08.720 --> 0:05:11.280 have to do? It seems like you could basically come 0:05:11.360 --> 0:05:14.280 up with a simple list of instructions like take drink orders, 0:05:14.400 --> 0:05:18.040 bring drinks, take the food order, bring the food, check 0:05:18.080 --> 0:05:21.600 the happiness levels at the table, bring the check, uh, 0:05:21.839 --> 0:05:25.840 process payment, and then you're done. But actually being a 0:05:25.880 --> 0:05:31.680 server in a restaurant requires hundreds of constant little improvisations 0:05:32.640 --> 0:05:36.000 h plus more athletic nimbleness than any robot today is 0:05:36.040 --> 0:05:38.440 even close to capable of. I mean, can you can 0:05:38.480 --> 0:05:42.159 you imagine with how clumsy robots are now, something that 0:05:42.200 --> 0:05:45.640 could like clean tables and move quickly back and forth 0:05:45.680 --> 0:05:48.520 between a kitchen and a table without running into people. 0:05:48.880 --> 0:05:51.760 Keep in mind also that most robots are very very 0:05:51.800 --> 0:05:55.479 good at navigating static environments. So in other words, you 0:05:55.520 --> 0:05:59.719 program in the knowledge of what the environment is, you 0:05:59.760 --> 0:06:01.960 know how it is, and give the robot some ability 0:06:02.000 --> 0:06:05.560 to sense its environment. It can then navigate around obstacles. 0:06:05.800 --> 0:06:08.680 But in a place like a restaurant, you have a 0:06:08.760 --> 0:06:11.839 constantly shifting environment. You've got people getting up and leaving, 0:06:11.880 --> 0:06:15.679 You've got chairs that are moving. So you have a 0:06:15.800 --> 0:06:19.200 robot that has to have really good path finding technology 0:06:19.279 --> 0:06:23.040 to maneuver its way through a constantly shifting environment. That's 0:06:23.120 --> 0:06:26.400 really hard to do. Sure, but there's another factor that 0:06:26.400 --> 0:06:31.640 we haven't considered yet, which is, you know, in defensive servers, 0:06:31.720 --> 0:06:34.800 we know that you guys are not robots. Yeah, there's 0:06:34.839 --> 0:06:38.159 other stuff that you're doing than just taking the orders 0:06:38.160 --> 0:06:41.120 and bringing the stuff exactly right. And I think this 0:06:41.200 --> 0:06:44.320 might actually be the biggest problem with the idea of 0:06:44.320 --> 0:06:48.320 a robot server, which is that servers need social intelligence 0:06:48.760 --> 0:06:51.840 good ones do. Yeah, well, this might be the biggest 0:06:51.839 --> 0:06:54.279 issue of all here, because a server in a restaurant 0:06:54.360 --> 0:06:58.520 isn't just a food delivery machine. A good server makes 0:06:58.560 --> 0:07:02.839 people feel welcome, makes and feel comfortable, uses charm to 0:07:03.040 --> 0:07:07.400 like up sell drinks and specials and appetizers, can explain 0:07:07.520 --> 0:07:10.760 things about the restaurant and answer questions about the menu, 0:07:10.960 --> 0:07:14.880 explain things about the food, describe what something tastes like, 0:07:15.040 --> 0:07:20.400 and then especially can sort of like calm complaints and 0:07:20.440 --> 0:07:23.800 help people get what they want, understand what the people want, 0:07:24.120 --> 0:07:29.240 and make them feel happy. Like can you imagine, let's 0:07:29.240 --> 0:07:31.800 say you're out with your family and you know your 0:07:31.840 --> 0:07:35.360 great aunt she has a complaint about her chicken peccata. 0:07:35.680 --> 0:07:40.160 Can you imagine a robot server being able to make 0:07:40.200 --> 0:07:42.920 her feel happy and like make her feel like her 0:07:42.960 --> 0:07:46.440 complaint had been properly dealt with. I can't imagine anyone 0:07:46.520 --> 0:07:50.040 making my grain on feel happy. Well well, but especially 0:07:50.080 --> 0:07:53.120 not the same kind of software that like answers your 0:07:53.240 --> 0:07:58.240 telephone call you call the insurance company, and like that 0:07:58.320 --> 0:08:00.640 thing is what's trying to like make you fear good. 0:08:00.760 --> 0:08:04.520 Or maybe it's the chatbot that only answers answers any 0:08:04.640 --> 0:08:09.320 any message with a question I'm not happy. Why do 0:08:09.360 --> 0:08:13.640 you think you're not happy? The doctor? Yeah. So, in 0:08:13.680 --> 0:08:16.120 the context of robots taking our jobs, I think this 0:08:16.280 --> 0:08:20.760 also means that, in addition to creativity and improvisation, another 0:08:21.240 --> 0:08:25.160 type of job that will be safely human for a while, 0:08:25.280 --> 0:08:29.120 our jobs that have a strong demand for good social intelligence. 0:08:29.880 --> 0:08:32.720 So what does social intelligence mean, I mean, we we 0:08:32.760 --> 0:08:34.640 sort of gave an idea there, but how would we 0:08:34.679 --> 0:08:37.760 define it? I would say one thing is that it 0:08:37.800 --> 0:08:41.320 means something along the lines of being able to read 0:08:41.400 --> 0:08:46.400 social cues, to perceive and understand emotional states, and to 0:08:46.559 --> 0:08:51.240 intentionally manage and trigger targeted emotions in others. Right, In 0:08:51.240 --> 0:08:53.520 other words, in order that it has to be able 0:08:53.600 --> 0:08:58.200 to navigate social situations in a way that appears to 0:08:58.240 --> 0:09:02.320 be natural and does not make things worse, right, doesn't 0:09:02.320 --> 0:09:06.880 make someone feel awkward or embarrassed or ill at ease 0:09:07.040 --> 0:09:11.160 for any reason, whether because you are communicating with a 0:09:11.200 --> 0:09:16.439 and obviously artificial construct, which already for many of us, 0:09:16.800 --> 0:09:20.680 is uncomfortable. Right, Yeah, some of us We we in fact, 0:09:20.880 --> 0:09:24.080 humans are pretty flexible creatures. We can adapt fairly quickly 0:09:24.120 --> 0:09:27.040 if we're willing to, sure, sure, And and some of 0:09:27.120 --> 0:09:29.920 us I mean, I suspect many people listening to this podcast, 0:09:29.960 --> 0:09:33.680 for example, probably find it charming that the novelty of 0:09:33.720 --> 0:09:36.920 that kind of interaction, although really like like after a while, 0:09:36.960 --> 0:09:40.240 you'd be like, I just want not burned chicken cottage, 0:09:41.600 --> 0:09:45.560 And some of us might actually prefer the interaction with 0:09:45.600 --> 0:09:48.600 a robot as opposed to a human being. That's true. 0:09:48.720 --> 0:09:50.560 That's true. I have days where I'm not sure that 0:09:50.600 --> 0:09:52.680 I can do this list of things that we're talking 0:09:52.720 --> 0:09:56.480 about here that has grouped under social intenation. That's true. Well, 0:09:56.480 --> 0:09:58.640 but I mean we all have to practice our social 0:09:58.679 --> 0:10:01.360 intelligence skills all the time. Oh sure, yeah, no, this 0:10:01.360 --> 0:10:03.800 this is not something that we think. It's not something 0:10:03.840 --> 0:10:06.400 even humans are always perfect at. Right now, there are 0:10:06.400 --> 0:10:08.560 plenty of times where you might be wrapped up in 0:10:08.760 --> 0:10:11.240 something that's going on in your life and you don't 0:10:11.240 --> 0:10:14.160 pick up on social cues that otherwise you would notice 0:10:14.280 --> 0:10:17.160 right away. Um, there are plenty of times. I mean 0:10:17.280 --> 0:10:21.120 I certainly have been guilty of that because I'm I'm 0:10:21.200 --> 0:10:25.360 always wrapped up in my own mind. Uh So, you know, 0:10:25.400 --> 0:10:27.880 when we're when we're aware, and when we are capable 0:10:28.000 --> 0:10:31.079 of seeing it, then that's one thing. But machines, they 0:10:31.080 --> 0:10:34.000 don't have that innate ability at all. Right, you have 0:10:34.080 --> 0:10:37.840 to build that into machines. And now there is an 0:10:38.040 --> 0:10:43.960 entire emerging field of artificial intelligence that is dedicated to 0:10:44.200 --> 0:10:48.480 this problem, to the problem of helping machines understand human 0:10:48.480 --> 0:10:52.360 emotional states and manage them as best as possible, and 0:10:52.440 --> 0:10:54.600 sort of all the other things we think of as 0:10:54.880 --> 0:11:00.000 social intelligence, you know, impression management and and navigating complex 0:11:00.080 --> 0:11:04.559 social scenarios, not being annoying, you know, all these weird 0:11:04.600 --> 0:11:08.080 things that they don't come naturally to machine. Right. And 0:11:08.080 --> 0:11:10.560 and you know, this is important not just for robots 0:11:10.600 --> 0:11:15.040 that are specifically designed to have social interactions, but robots 0:11:15.040 --> 0:11:18.760 that are just going to be around humans. You know, 0:11:19.040 --> 0:11:23.599 they're especially consumers, right because you know, people robot industry 0:11:23.640 --> 0:11:26.840 people are kind of used to having to program a 0:11:26.920 --> 0:11:30.160 robot by typing code into a thing, you know. But 0:11:30.160 --> 0:11:34.520 but consumers, Yeah, and it kind of and it may 0:11:34.559 --> 0:11:38.160 be that the robots that you, as a consumer encounter, 0:11:38.240 --> 0:11:40.880 it may be that there's no direct interaction, but they 0:11:40.920 --> 0:11:44.160 still those robots still need to have that social intelligence 0:11:44.160 --> 0:11:47.480 to understand how to navigate through while being as unobtrusive 0:11:47.520 --> 0:11:50.480 as possible. It's it's one of the many things that 0:11:50.559 --> 0:11:53.320 we have to take into consideration as robots take an 0:11:53.360 --> 0:11:58.240 increasingly more prominent role in our lives, along with the 0:11:58.280 --> 0:12:00.439 idea that the robots have to be design in such 0:12:00.440 --> 0:12:03.640 a way that they are not likely to cause harm 0:12:03.760 --> 0:12:06.720 to people through their normal operations. I mean, this is why, 0:12:07.160 --> 0:12:10.280 like we've talked about before, with big industrial robots. Typically 0:12:10.320 --> 0:12:15.080 there are huge safety rails all around them, because these 0:12:15.120 --> 0:12:18.959 are just machines that do the same set of actions 0:12:19.080 --> 0:12:21.640 over and over again. And if you let a team 0:12:21.640 --> 0:12:24.280 full of preschoolers run around and you get a bunch 0:12:24.280 --> 0:12:27.520 of preschoolers welded to a metal wall, it's not good. 0:12:28.120 --> 0:12:32.080 You don't want that. So this we're we're talking about 0:12:32.120 --> 0:12:36.480 the social equivalent of that same thing. You don't You 0:12:36.520 --> 0:12:39.640 don't want the social equivalent of welding preschoolers to a wall. 0:12:40.280 --> 0:12:42.200 I'm not sure what that looks like, but I kinda 0:12:42.320 --> 0:12:45.839 I kind of want. Can someone please make us I'll 0:12:45.880 --> 0:12:48.320 take you to my next family reunion. You will be 0:12:48.360 --> 0:12:52.440 able to experience it. No here, I think here's the 0:12:52.440 --> 0:12:55.840 the real world takeaway. We're gonna have robots in our homes. Yeah, 0:12:55.960 --> 0:12:58.520 I mean, that's that's pretty clear. We aready. Yeah, we 0:12:58.600 --> 0:13:01.920 already might have some roomba or something like that. More 0:13:01.960 --> 0:13:06.600 and more we're gonna be incorporating robotics and artificial intelligence 0:13:06.679 --> 0:13:10.800 into our home life. And it actually matters that these 0:13:10.840 --> 0:13:15.920 things are not causing emotional stress and annoyance, like making 0:13:15.960 --> 0:13:21.520 our lives more unhappy by not knowing how to behave themselves. Yeah, 0:13:21.520 --> 0:13:24.560 it's a great point. And so there are entire research 0:13:24.600 --> 0:13:28.800 departments dedicated to this particular aspect of artificial intelligence, And 0:13:28.840 --> 0:13:30.839 in fact, there's one that's right down the road from it, 0:13:31.000 --> 0:13:33.280 right here at Georgia Tech. They've got a socially Intelligent 0:13:33.360 --> 0:13:36.000 Machines Lab. Yeah. We'll talk about one of the robots 0:13:36.040 --> 0:13:38.600 that they've worked on a little bit later, but I've 0:13:38.600 --> 0:13:41.959 had a chance to visit that particular lab. Yeah, it's 0:13:42.120 --> 0:13:45.760 pretty darn cool. Guess. Don't mean to brag, but it's 0:13:45.760 --> 0:13:49.360 pretty neat. Even as even as a University of Georgia graduate, 0:13:49.400 --> 0:13:54.000 I can appreciate this particular Georgia Tech lab. Uh. So 0:13:54.760 --> 0:13:58.240 let's talk a little bit about what it means for 0:13:58.520 --> 0:14:01.800 robots to actually possess social intelligence. You touched on this, Joe, 0:14:01.840 --> 0:14:04.480 you were talking about the ability to pick up on 0:14:04.480 --> 0:14:07.480 those social cues. So I would I would say that 0:14:07.720 --> 0:14:11.360 a socially intelligent machine has to be able to observe, analyze, 0:14:11.360 --> 0:14:15.079 and respond to humans accurately and within the proper context 0:14:15.280 --> 0:14:19.480 of the situation. So one example, and I've seen this 0:14:19.600 --> 0:14:24.360 as particularly examples in Japan, robots that are part of 0:14:24.440 --> 0:14:27.680 a hotel check in phase. You come in and you 0:14:27.680 --> 0:14:30.720 you need to get your room. You would want that 0:14:31.080 --> 0:14:35.640 experience to be efficient and pleasant and probably not reliant 0:14:35.680 --> 0:14:39.000 on too many assumptions because that could get socially awkward 0:14:39.040 --> 0:14:43.200 depending upon what the nature of the hotel visit could be. Right, Sure, 0:14:43.240 --> 0:14:48.560 you wouldn't want, for example, you know, yeah, sir, that 0:14:48.680 --> 0:14:51.600 you have a lovely daughter, this is my wife. That 0:14:51.640 --> 0:14:55.200 could be awk kind of thing. Yes, that's that's one 0:14:55.520 --> 0:14:58.920 of the team or examples I could get, but yes, 0:14:59.680 --> 0:15:02.760 but in context is really important. So let's say here's 0:15:02.760 --> 0:15:05.520 another example. I've actually, you know, you we've talked about. 0:15:05.840 --> 0:15:07.680 I'm not sure if we've talked about on this podcast, 0:15:07.720 --> 0:15:11.680 but there was a design for a flying drone that's 0:15:11.720 --> 0:15:15.360 meant to encourage joggers to run. It was to keep 0:15:15.360 --> 0:15:17.880 pace with joggers so that they have they feel like 0:15:17.880 --> 0:15:20.480 they have a running buddy. So you know, Now, that's 0:15:20.480 --> 0:15:22.360 a simple version of a robot. But imagine that we 0:15:22.400 --> 0:15:25.680 have a more advanced one that actually gives verbal encouragement 0:15:25.800 --> 0:15:28.760 so that you can continue to run. Now, let's imagine 0:15:28.760 --> 0:15:31.520 that said robot veers off the path and into a 0:15:31.560 --> 0:15:35.200 memorial service at a cemetery. I think saying you can 0:15:35.280 --> 0:15:38.000 do it, get on up there, keep on moving at 0:15:38.000 --> 0:15:40.800 a memorial service that, Yeah, that might might not go 0:15:40.920 --> 0:15:43.760 over so well in the social situation. That is a 0:15:43.800 --> 0:15:47.720 memorial service. So when we're talking about social intelligence, we're 0:15:47.720 --> 0:15:51.640 talking about not just responding within the situation the robot 0:15:51.680 --> 0:15:55.720 was intended for, but in broader situations. Sure. And I 0:15:55.760 --> 0:15:58.640 think also side note, this means that we need to 0:15:58.680 --> 0:16:02.720 teach all robots in the future about graveyards and funeral practices. 0:16:02.760 --> 0:16:06.640 I think why they're cool. Yeah, now I'm just seeing 0:16:06.720 --> 0:16:10.400 a bunch of robots listening to the Smiths, and the 0:16:10.560 --> 0:16:13.440 robots says, hey, baby, let's go hang out in the graveyard. 0:16:14.160 --> 0:16:16.280 I'm feeling better and better about the lyric I picked 0:16:16.280 --> 0:16:19.120 for the beginning of this episode. Okay, So, in addition 0:16:19.200 --> 0:16:24.040 to understanding context, I think another crucial part of social 0:16:24.120 --> 0:16:28.840 intelligence is the ability to incorporate new information right because 0:16:29.040 --> 0:16:33.880 to right, Like, when we interact with people, we don't 0:16:33.920 --> 0:16:38.160 treat everybody the same way. I mean because they not 0:16:38.240 --> 0:16:40.520 everybody wants to be treated the same way you treat 0:16:40.560 --> 0:16:43.600 everybody else. You learn different ways to interact with different 0:16:43.600 --> 0:16:46.960 people based on what your relationship with them is, and 0:16:47.000 --> 0:16:51.600 what their preferences are. Some people like very jokey interactions, 0:16:51.640 --> 0:16:54.480 some people are a little more business like. Yeah, there's 0:16:55.400 --> 0:16:58.840 entire fields dedicated to this as well. And it's interesting 0:16:58.880 --> 0:17:01.920 because the law of the earning behaviors I've seen have 0:17:02.000 --> 0:17:06.080 concentrated first on something a little simpler, because when we 0:17:06.080 --> 0:17:10.040 start getting into social interactions, that's a really complex and 0:17:10.240 --> 0:17:15.000 and chaotic field. So, I mean, if you feel stress 0:17:15.080 --> 0:17:18.359 in social situations like social anxiety, a lot of the 0:17:18.400 --> 0:17:21.439 stress you're probably feeling is it has to do with 0:17:21.480 --> 0:17:24.480 the complexity of the social interactions that we have. In 0:17:24.560 --> 0:17:28.200 your fear that you're not meeting expectations, writing, unsure that 0:17:28.440 --> 0:17:32.080 the choices that you're making are are good choices. Right, yeah, now, 0:17:32.640 --> 0:17:37.480 and that's a great point. Now, obviously the robots aren't 0:17:37.480 --> 0:17:40.639 going to be experiencing social anxiety, but they could end 0:17:40.760 --> 0:17:43.760 up that we know of, but they could certainly encourage 0:17:43.760 --> 0:17:45.840 it in other people if they the robots are not 0:17:45.880 --> 0:17:48.520 behaving in a way that the people were expecting. If 0:17:48.560 --> 0:17:51.040 they have said something and the robots seems to behave 0:17:51.080 --> 0:17:54.440 in a in a completely counterintuitive way, that could really 0:17:54.480 --> 0:17:56.200 cause some issue. Oh yeah, well, I was just trying 0:17:56.200 --> 0:17:59.719 to communicate the complexity like, Yeah, us humans who are 0:17:59.720 --> 0:18:02.760 pretty good at social interaction have enough trouble with it 0:18:02.760 --> 0:18:06.480 that sometimes we can get anxiety about it. How So, 0:18:06.680 --> 0:18:10.400 a lot of the learning I've seen has been oriented 0:18:10.440 --> 0:18:14.320 to teaching a robot to perform a series of tasks 0:18:14.600 --> 0:18:17.720 by performing them for the robot. Actually the one that 0:18:18.040 --> 0:18:20.560 there's one in Georgia Tech that does this where you 0:18:20.600 --> 0:18:24.080 could have a robot and it goes into observation mode 0:18:24.080 --> 0:18:26.560 and watches what you do, so you could pick up 0:18:26.640 --> 0:18:28.640 let's say that you've got, you know, kind of going 0:18:28.680 --> 0:18:31.160 back to that example of the sphere in the box. 0:18:31.240 --> 0:18:32.800 Let's say you've got a sphere on a desk in 0:18:32.880 --> 0:18:35.760 a box in the desk, and you open up the box, 0:18:36.160 --> 0:18:38.600 you pick up the sphere, you put the sphere in 0:18:38.720 --> 0:18:42.359 the box, you close the box. The robot will observe 0:18:42.440 --> 0:18:45.520 this and then be able to repeat those steps. And 0:18:45.600 --> 0:18:48.480 even if if you've designed the robot properly, it might 0:18:48.520 --> 0:18:52.560 be able to ask questions like does it matter what 0:18:52.840 --> 0:18:55.879 how I pick up the this thing? Does that matter? 0:18:56.160 --> 0:18:58.640 Does it matter how I close the box doesn't? If 0:18:58.640 --> 0:19:00.840 it's four flaps, does it matter? Or which flaps are 0:19:00.840 --> 0:19:03.320 closed first? Is there a sequence? Right, if it were 0:19:03.320 --> 0:19:05.240 a cube that you were putting inside the box instead 0:19:05.280 --> 0:19:07.520 of his sphere, doesn't matter which side of the cube 0:19:07.520 --> 0:19:11.679 is facing up perfect exactly right. So that's that's a 0:19:11.720 --> 0:19:15.960 building block, not to make it a weird, stupid pun, 0:19:16.040 --> 0:19:19.520 but that's a building block for the kinds of interactions 0:19:19.560 --> 0:19:22.840 where a robot starting to get into more complex situations, 0:19:22.880 --> 0:19:28.080 such as, this person is exhibiting behaviors that indicate that 0:19:28.119 --> 0:19:31.000 they are in a happy mood. So what is the 0:19:31.040 --> 0:19:34.800 appropriate response for this particular person when they are in 0:19:34.840 --> 0:19:39.000 a happy mood. Maybe it's playing that person's favorite upbeat 0:19:39.080 --> 0:19:43.159 song because it's going to enhance that happy mood. And 0:19:43.200 --> 0:19:46.920 then that's what the robot ends up doing. Uh, these 0:19:46.960 --> 0:19:49.960 are you know, it's going to take these these basic 0:19:50.080 --> 0:19:53.640 steps before you get into something so complex. Is all right, 0:19:53.640 --> 0:19:57.640 Now we've got two people, two people the robot has 0:19:57.680 --> 0:19:59.960 to interact with. If we're talking about a home robot, 0:20:00.000 --> 0:20:02.639 it might have profiles for each member of the family, 0:20:02.720 --> 0:20:06.080 for example, but it knows that person number two is 0:20:06.080 --> 0:20:09.199 not really into music, so playing songs aren't going to 0:20:09.359 --> 0:20:13.040 have that same emotional effect as for the first person, 0:20:13.400 --> 0:20:15.600 So it may have to build an entirely different set 0:20:15.640 --> 0:20:18.680 of interactions, just as we humans would kind of intuitively 0:20:18.800 --> 0:20:21.919 know from our interactions, like what you know works and 0:20:21.960 --> 0:20:24.760 doesn't work with the people in our lives. And you 0:20:24.800 --> 0:20:28.600 start to realize, Wow, this is this is tough. This 0:20:28.680 --> 0:20:35.359 is a non trivial challenge in artificial intelligence. Absolutely. Let's 0:20:35.400 --> 0:20:38.280 talk about some of the teams and some of the 0:20:38.359 --> 0:20:42.639 robots that are that are taking these challenges head on. Sure. Yeah, 0:20:42.760 --> 0:20:45.760 so we've got a collection of different robots. These are 0:20:45.800 --> 0:20:49.639 These are mainly robots designed specifically to have social interactions 0:20:49.640 --> 0:20:52.640 with humans, and lots of we should say that lots 0:20:52.720 --> 0:20:57.760 of computer programs and robots have some elements of this. Yeah. Um, 0:20:57.800 --> 0:21:00.359 I mean typically you're not going to be interacting with 0:21:00.440 --> 0:21:04.280 a robot that's just designed to be social, because what's 0:21:04.320 --> 0:21:07.080 the point other than a novelty. I mean, typically social 0:21:07.119 --> 0:21:11.800 intelligence is a feature of a robot that intended to 0:21:11.880 --> 0:21:15.440 do something else. Yeah, except for I guess therapy, but yeah, 0:21:15.480 --> 0:21:17.920 that's true. Therapy, butts would be a great example. Yeah. 0:21:17.920 --> 0:21:19.960 I started to think that in some of these, some 0:21:20.040 --> 0:21:22.720 of these cases, we're heading toward a future like in 0:21:22.800 --> 0:21:27.840 future rama, where we have robots built to be hoboes 0:21:28.280 --> 0:21:31.560 or gamblers or whatever, and you think, why would anyone 0:21:31.600 --> 0:21:35.720 ever build a robot, But we are doing that. Yeah. 0:21:35.840 --> 0:21:39.119 So one robot that I actually watched this adorable video 0:21:39.440 --> 0:21:42.920 earlier today was this robot called Gibo. Had you heard 0:21:42.960 --> 0:21:45.840 about this one before? I have? I had seen pictures. 0:21:45.880 --> 0:21:48.200 I had not seen the video or read about Gibo, 0:21:48.320 --> 0:21:50.840 but I had seen photos of Gibo before. Okay, so 0:21:50.880 --> 0:21:54.720 this is nothing super complex. Gibo is just supposed to 0:21:54.800 --> 0:21:57.160 be a social robot that's really designed as a kind 0:21:57.200 --> 0:22:00.480 of around the house helper. So imagine kind of an 0:22:00.480 --> 0:22:04.879 embodied serie in a way. It can track faces and 0:22:05.080 --> 0:22:11.200 recognize individual household members, take pictures, record voice reminders, um 0:22:11.320 --> 0:22:13.320 and do a lot of the stuff you might have 0:22:13.359 --> 0:22:15.840 like apps on your smartphone. It can even do things 0:22:15.920 --> 0:22:19.520 like um like sort of like a similar feature as FaceTime, 0:22:19.680 --> 0:22:23.040 where you are in one place, one of your family 0:22:23.080 --> 0:22:25.960 members is somewhere else they are using their smartphone, and 0:22:26.000 --> 0:22:29.479 then Gibo just projects their face onto its face. So 0:22:29.560 --> 0:22:32.880 you have a a direction to look in and it's 0:22:32.880 --> 0:22:34.960 an actual physical thing. It's not something you're holding in 0:22:34.960 --> 0:22:38.120 your hand. You're looking down at Gibos quote unquote face, 0:22:38.119 --> 0:22:40.680 which is really kind of a round screen. Yeah, it's 0:22:40.720 --> 0:22:44.120 this like it's this kind of friendly. I yeah. And 0:22:44.160 --> 0:22:45.960 it has a camera in it so it can it 0:22:45.960 --> 0:22:49.200 can quote unquote see that allows it to do the tracking. 0:22:49.240 --> 0:22:51.320 Now it doesn't like walk around the house or anything. 0:22:51.359 --> 0:22:53.600 It's sort of on a stationary base. But you could 0:22:53.680 --> 0:22:55.840 pick it up and move it around with you a 0:22:55.840 --> 0:22:58.879 little more than five pounds, I think. And the thing 0:22:58.920 --> 0:23:02.200 about this is nothing about this robot is like crazy 0:23:02.280 --> 0:23:05.920 mind blowing, you know. It's not like doing anything all 0:23:05.960 --> 0:23:10.360 that weird. It's sort of a cuddly smartphone or cuddly 0:23:10.560 --> 0:23:14.000 series kind of thing. It you can talk to it, 0:23:14.000 --> 0:23:17.359 it'll talk back to you. It tries to, uh to 0:23:17.600 --> 0:23:20.960