 likable science on a given Tuesday. I'm Jay Fidel. This is Think Tech and with me is Ethan Allen. He and I have been doing likable science. It's his show actually for years now and we find that science is likable and we also stand in awe of it. And today we're talking about robots because if you didn't know robots are really really going places. Welcome to your show Ethan. Nice to see you. Good to be here. It's good to see you again. So yeah robots. I mean I was telling you that I was in high school or junior high school. I had a play. I studied a play called Rusman's Universal Robots and it was written by some guy in Eastern Europe and it was about a community that had designed robots and in fact in fact they say that the word robots came out of this play. It was invented in this play. We come a long way. This was in 1920 or so and people were fascinated with science then as they are now. But we've gone a long way and we have now nation states that are determined to be ahead of the game and be number one. So who is number one? Do you have an idea? I really don't. I mean it would really I think depend on sort of what type of robot you you talk about. Certainly the Japanese have done some remarkable sort of friendly robots. Sort of household I don't want to say exactly pets household companions and helpers. And I know they've produced quite a slew of them. Some just really very much directed to almost provide emotional support. Other is much more sort of just a domestic servant as it were. Yeah well they say that robots really are wedded somehow with the AI artificial intelligence technology. But you know I think we need to explain that to people. What is the connection? Why is AI so valuable in developing robots? Do you have any thoughts on that? So AI really covers a whole set of technologies having to do with what sometimes called machine-based learning or sometimes deep learning where traditionally all of our computers had to be programmed step by step by some person who understood the process and had to really lay out for the computer do a then do b then do c in a very structured process. Some while ago they as we got more and more sophisticated in our computer architecture they began developing machines which didn't need that level of instruction anymore. These machines can look at a problem and start figuring it out on their own. So one of my favorite examples is the game of Go which is a beautifully simple oriental game if you know it just placing little colored stones on intersections and trying to surround your opponent's stones while they're trying to surround yours. Very simple game very subtle and sophisticated. The standing saying about Go is you can learn it in five minutes and spend the rest of your life learning to play it well. And some years ago AI machines began beating the world's best Go players that they would literally play millions of games against themselves and figure out completely new strategies that people never had thought of so but translating that into a robot which is much more we think of as a machine that does some task some physical work for us either picks up things and carries them for us or delivers us some drugs or packages or mail or what have you or you know it's much more a physical presence. So naturally you can see there's this tendency to want to put them together and say wow let's put some of that artificial intelligence into one of these machines and then it would be a much more useful companion you know it would know it could sort packages much more efficiently and deliver them more cleverly it could take a whole bunch of different things and deliver them to the right places in the right sequence do it quickly and efficiently. There are certainly lots lots of powerful ways these things can come together but they are really very sort of separate concepts. Yeah well I make a list in my mind you know thinking of all the different kinds of robots we have seen you know in relatively recent years I mean take robotic planes the drones they're large you know you can control them from a bunker somewhere in in Colorado but you can also set it on its own path and and undersea robots the same way where it knows what its mission is and it just does that and it's capable of adjusting to changes in circumstances and they're getting smarter and smarter I mean you know I think 10 years ago they were not that smart but they're much smarter now and then you have the vacuum cleaners at home and you have I would tell you before the show this was recently I sent it to you it was either on YouTube or it was on one of my television shows about technology it was a story of an American company where their mission was to develop robots that would actually engage with humans it's sort of a hybrid where humans were better at certain kinds of functionality robots were better at other kinds and if they could only work together and so the robot was designed to get out of the humans way so it was a workstation kind of concept the robot would help the human and get out of his way or her way when he needed her to move aside and I didn't know that was something I'd never seen before because usually we think of the robot just do it and we don't interfere with the robot but now it's changing now it's more humanistic somehow right one of the big places I think we've seen a lot of robotics just really quite recently is is in cleaning in workplaces and hospitals and all and they have robots now that can be once they're sort of figure out a room can go around and clean the room but they have to do that with some people are there people are moving in and out walking around and the these unlike your little room but which basically you know your little home vacuum will basically run into your feet right and then it jumps away move somewhere else but it would come in right back and hit your feet again these apparently will avoid people very nicely which is good because you don't want you don't want things come up and spray antiseptic all over you necessarily so yeah and your earlier examples of the drones and underwater unmanned vehicles are really nice examples of where that intersection of artificial intelligence and sort of robotic technologies come together in very powerful ways now so you know I was thinking the other day there's that you know recently I saw the movie 1917 yeah and a lot of that is the battlefield where you know human human people human sacrifices were made where thousands hundreds of thousands even millions of people were thrown at each other you know to make to make to kill and it strikes me that you know we have we have come a long way we have the atomic bomb which is not a good thing because that kills a lot of people and we have certainly these drones which are which in their own way are much better than planes because there's no risk to the pilot and then of course the battlefield itself you know you see all these battlefield situations all over the world and you say gee what why don't I just make a bunch of robots it reminds me of you know in China the clay soldiers in the west of China Xinjiang I want to say just hordes of soldiers but they're all robots and you go out of the battlefield and do what they're supposed to do they have their own mission and it's out of science fiction but you know I think that's got to be coming down the pike if you want to avoid risk to a human being make a robot if you want to achieve some larger mission which is repetitive tasks or tasks that we can handle by AI make a robot make a lot of robots you know and then they have these swarms lately have you been following the swarms right yes the small small drones that behave as as swarms pay attention to one another and essentially gain collective intelligence from working together right yeah potentially very very great potential as weapons but that's a huge matter of huge debate among you know military types now is do you turn that final decision over go no go you know set off the explosive don't set it off do you turn that over to the machine let the machine make that decision entirely without human direction you point out you can send these drones on their merry way and go off and do something else and the drone will fly around avoid obstacles swing around storms go over the mountains whatever it needs to do and get to its target leaving in the battlefield leaving that level of control out of human hands is a very interesting moral dilemma I think I think the generals and admirals have considerable debate about this because it's it has truly awful potential to go wrong right oh absolutely I mean you could design a whole fleet that would be you know automated and yet every every device you design can be hacked also so you send the fleet north and somebody hacks it and now the fleet's coming south this is not exactly what you wanted exactly I'm particularly if that fleet is armed with nuclear missiles or whatever right it has potential to get very ugly very fast and I think the military's my understanding military's been fairly reluctant to take that step of turning stuff over they'll use them they'll use drones widely for sensory stuff they'll deliver uh can deliver it but they have tended so far to keep a human finger on the button that were to make that final decision yeah oh and yeah there was a recently also automated trucks american company making automated trucks and they cross country without a driver and although they have safety drivers now in them um safety driver is only there for safety and in fact the truck can get across the country without any any human intervention uh of course the you know the long-haul drivers are not particularly happy about that but the fact is it's a lot cheaper to ship goods that way and it's a lot safer the truck the truck doesn't fall asleep the truck doesn't you know drink too much and so forth yeah I mean if you consider the autonomous vehicles self-driving vehicles which indeed are are a kind of robot right they are in their very primitive crude stage right now they're we're really just we're just beyond prototype more we're field testing you know model A's model T's you know and yet their per mile driven fatality rate is less as I understand less than a quarter that of humans at this point um that is they are essentially three times safer than than having human drivers and this is just you know first generation crude stuff that's making mistakes you know one can only imagine that numbers that ratio is going to improve these things are essentially going to become I think so much safer that that people realize we have to we have to let leave that that's a task that really robots can do much better than people fundamentally yeah now the one the one thing in this in this program that touched me was one of the drivers who opposes the long-distance robotic trucks said well suppose there's an accident and there's a policeman the truck was not involved in the accident but there's a policeman who says everybody off the road I want I want to clear the road so we can remove you know the vehicles that were involved in the accident well the truck wouldn't know the truck would have no idea you know what the cop was saying and so you know you have a problem because the truck is programmed to do this but not that I say to myself well this goes back to that whole thing about respecting humans of having a you know a robotic device respecting humans so if the robotic device you know notices as a policeman and he's giving hand signals the truck reports back to some some person in a bunker in Colorado say for example and and and he takes over and he's driving the truck and off the road whatever the cop wants because he can see what the cop is asking and he can hear so I think in the end it's it's hybrid you know you have to have one working with the other right I mean you know you what you were saying there is really reminiscent of you know Asimov's three laws of robotics right you know that is a gas monster of rota as the necessary precursor to really letting robots loose letting them do their things is they they have to have that built in that they cannot hurt people you know because otherwise otherwise the whole situation gets pretty grim well the same time you know that we have robotic trucks we're also having computers that with AI that think and learn and you know can determine their mission or redefine their mission even and you know I don't know if anybody would ever claim that anything right now today is capable of consciousness but that's coming I mean it's not going to be the same consciousness that you and I have but it's kind of a consciousness it's coming well you know there are debates going on now about sort of the the ethics of artificial intelligence and the sort of the morality of machine learning us or and people are considering these because we have to we're going to face it we're going this is going to come upon us it is sort of like some of the biotechnology currently that we have these we have the capability right now for instance as has been demonstrated to edit the human genome and produce viable people who are gene edited is that okay to do the general scientific consensus is no we really should not be doing that at this point and very similarly with with artificial intelligence and robotics we need to have some very serious discussions about what is okay and what isn't okay and then sort of get everybody to agree to that set of rules and to sort of agree to play fair by it uh which is probably the trickier part well you know and then the thing about it talking about hacking right uh hacking my fleet going north and make it go south that's not that's not going to stop i mean if you we did a show a day or two ago about about uh the you know the new scams that have come up since covid and some very creative smart people you know around the world a lot of them are nation states actually that fund things like like russia does this kind of thing you know with the uh internet research agency you know and hacking our elections would have you but there's a lot of people a lot of nation states a lot of individuals a lot of a lot of organizations around the world that that hack and you're not going to be able to stop them not not in the current you know paradigm and and so if if we design uh robots which are really smart and almost conscious um there's always going to be a back door there's always going to be some hacker who wants to get in there and redirect that robot for his own purposes and that's pretty scary that's never going to be able to close up we're never going to be able to lock the robots up so nobody can touch them nobody can redirect them don't you think they'll all be on the internet so part of the internet of things they'll all be directable or connectable by remote and they'll always be somebody who will want to intervene i think that's true i think that's that is human nature that you know the technologies themselves are all sort of value neutral it really is up to us as people to decide how those technologies get used and one would hope that the better nature of human beings you know tops wins out on in this case and that we use these technologies wisely and for general and human benefit given sort of our history with technology i i don't know that's a safe assumption to make that will happen yeah no i you know because um in fact you as we discussed you can have a a perfect um dictatorship if you can use data processing and and um you know ai to control your population it's coming i mean and you know it's no problem at all there's you know there's dozens of programs around that can handle billions of records and billions of fields a matrix is as big as you can ever imagine and if you have a population you want to know everything about that population you can get the data you can put the data in that database and wow you have a huge leverage over the population and it's very hard to unseat a dictator who uses that kind of technology right when you get multiple kinds of databases and have them linked up through artificial intelligence so you've got people's financial data their travel data their social history data their criminal data their uh food buying data their shopping data their daily routine data and you put all these together you know with a good smart algorithm can look and say hey joe there has now suddenly changed his pattern and he is you know leaving his house 20 minutes before he used to leave his house and he's walking a different route than he used to you know and immediately you know if you're a dictator you're gonna flag joe why why is he doing this you know and you'll send somebody to investigate because you know that that's a you know it's a it's a red flag so yeah it's it's it's truly frightening and and as robots become more ubiquitous of course they will likely be gathering as their sensors become better they'll likely be gathering that more and more of that same kind of data and it's unfortunate sure in Belarus you know they have these guys out there with with cell phone cameras and they're the idea is they're taking pictures of all the faces okay and of course that goes into facial recognition because those cell phone cameras are you know very powerful now and facial recognition is going to keep them in a database so you know who was protesting in Belarus this is pretty scary stuff but the next the next thing is with with that facial recognition and with you know physical robots and we've seen America a lot of American companies are building physical robots that can negotiate any terrain they look like animals you know and they have four legs and they can climb stairs and jump and and you push them and they and they they fall down but they come right back up again you've seen those photographs seems to me that it won't be long before we won't have to send the brown shirts into Portland or a given city we won't you know the federal government or any government can have brown shirts that are robots and they would go into a city and they can quell a disturbance or a protest or a demonstration even if it's first amendment you know in a matter of seconds because you can't do anything about it and you can't stop them and if they're on a mission to to quell that disturbance don't get in their way you will be hurt yeah I can see that happening that circles back to your earlier point is are we willing to build robots that would hurt people you know in the in the course of carrying out their mission and again one would hope not one would hope you would you would build robots that would have sort of built-in safeguards and would know at some very fundamental level of their of their consciousness of their programming but you know humans are sacred and not not messed messed with you know well we we're using robots now to um assassinate uh just you know it was only a year ago that you know robotic plane or a weapon of some kind assassinated that fellow in Iraq wasn't it and I think that's you know to the extent that there was a a moral barrier there I think we've we've passed that barrier in in in the name of doing the right thing but but it opens the door the genie is out of the bottle and anybody who feels he's doing the right thing can likewise get that I don't think that that technology is all that hard to get you know this kind of technology in some ways it may be controlled but in other ways it's it's out there and and anyone who you know for example in in Belarus that that dictator who claims he got 80 percent on the vote and we know that he didn't um if he wanted to quell the disturbance of that group of what is a million people there and in the center of but in Minsk in Minsk um and he had the robots to do it there's no question of my mind he would do it and if anybody got in the in the way of those robots uh he would have no reservation at all about about having the robots do do violence to those people uh I suggest to you this is coming don't you think either it's coming yeah and your point about this technology is becoming more both more ubiquitous and cheaper all the time and we've seen that trend I mean if you you look these days what what you spend to buy a good computer you're buying so much more power computing power than it was the impossible even spending 10 times that some 10 years ago you know we see this trend and so yes it's very likely that more and more people are going to get a hold of more and more powerful technologies uh that that sit at this interface of artificial intelligence and physical robotics and you know once that's out of the state government's hands it's uh it's in the hands of nefarious groups and wacky individuals who knows where it's going to go yeah think of a bank robbery I'd really rather not do a bank robbery I stand a good chance of getting shot caught imprisoned punished what have you but if I could get a robot to go do the bank robbery an anonymous robot you know that's not traceable I'd send that robot at any time right and how does it do it's getaway that's the problem yeah well it has to do with run you know I'm thinking of all those movies with the robots move with the wind but it's a confluence isn't it you know on the one hand the rabbi the robot can do a bank robbery on the other hand well you can you know you can do a hack too and you can steal from the bank on the internet and you know if you have the right tools right and so it all sort of comes together and the question is the morality point you raised and and if you're if you're a nefarious individual or a state um you know what are you going to do where are you going to stop where's the boundary and it's a moral question and I have a feeling that we're going to learn that moral barrier the hard way right and yet the you know the potential there is to have for real good too you were commenting on the robots that have this that have tremendous sort of physical abilities now to crawl into smaller spaces or whatever you know uh in humanitarian assistance in disaster relief settings those kind of robots are come incredibly valuable right because they can get in into collapsed buildings they they can sense uh through acoustics or whatever whether there are still people there where they are but it's between the person and the the exit so they have tremendous opportunities to use these kinds of technologies for real good one one can imagine in the wake of floods having robotic boats basically go around looking for people to rescue you know mapping that out telling other boats where they've seen people there's more people need you know directing a fleet of robotic boats coming around yeah fight fight fires fight fires that's a good one exactly in California right now yeah you know California could certainly use that um so you know again it's where I think we need I say we the powers that be really need to be discussing this the morality of robots and the the the ethical uses of artificial intelligence um and having some very serious discussions with not just the scientists but some of the end users philosophers religious scholars uh politicians you know get everyone involved because it's a very complex set of problems that really involves all of us ultimately if not today tomorrow or the next day it's going to see I think that's such a good point um you know uh we really haven't regulated for example drones the drones are still in a kind of tizzy um and nobody knows exactly what the limits are nobody enforces the limits and then you know the drones to drone operators to largely what they want um but it can be a very valuable thing to have a drone as you say for our you know crisis management incredibly valuable so what you need is government to regulate and unfortunate and and also make statutes of limitations so that you can't pass a certain line so that these moral questions are resolved by you know we're coming together of legislative bodies in a democracy unfortunately you know our ability to do that is is diminished right now in the time of covid and in this particular administration with the way congress is set up right now but in a in a perfect world what I hear you saying and I totally agree is that we ought to sit and reason about this we ought to sit and determine the morality we ought to make the boundaries we ought to regulate it intelligently not bureaucratically um and and and also fund the research all that exactly exactly you know um because you know the technology out there the pandora's box has been open the things are are coming out and people are doing great things you you may recall our friends at ocean it you know who hooked up a little camera with some a i and a drone and it can fly it over areas of storm and this drone can detect upright versus downed power poles and so just with a quick quick sweep can immediately send back the information on how many power poles aren't going to be needed how much power line is going to be needed where exactly it should be brought in how to get it there you know and really help restore you know recover from a disaster very quickly uh great stuff wonderful stuff you know so we're going to have fun great fun an extension of the of the human capability a friend if you will who helps us um get our work done makes life easier for us makes us more productive and you know what it just occurs to me here in the last minute of our show um is that is that we here in hawaii we could expert ourselves on this because it's a growing field it's a field where there's a lot of material out there to read and study there's a lot of collaboration possible um we have the beginnings of it we should we should make an effort uh to make it part of our reimagined economy don't you think absolutely absolutely it's going to going to be a growing area yeah well thank you Ethan it's always great to talk to you we always come up with things that we never anticipate we would come up with true we'll have to do it again soon i so enjoy having these discussions with you thank you so much Ethan thank you jay good to talk to you