A new story from Coast to Coast AM with George Noory


And welcome back George story with you Robert Elliott Smith with us our final segment Robert I've got to really applaud you for the work you did in your book rage because it's a very thorough thank you and I hope that is my goals were to be very accessible at the the book is about a complicated subject but what I'm trying to do is make it really transparent for everyone so I want people to to read the book if you don't have a mathematical background you don't you know with your computer scientists the books about getting everybody to understand hell simple well grins are held complex they are in our lives in light they're having such a big effect on us and so I really hope it's it's probably read did you have any more answers to all my yeah out there shall we move on I think I should say it's a little bit more she gave me this wonderful quote from a as men we don't see things as they are we see things as we are and and it's a great quote that statement work psychologically about that for the individual to say we see things through our own psychological life but it also works sociologically a lot of human intelligence is is basically social that we make decisions as creeps by informing one another and it allows us to deal with highly uncertain world that we live in think about that and think about that what I'm saying about social media if we're roped into communities that only communicate in in in groups that re self reinforce then we see things is that we had held out we is in there we then is very narrow that that group of people is very narrow in its sporty view and if people want to then attack how you see the world they know where to go they know the near community to get to and say things specifically to them so this is this is this dynamic I'm talking about and and there are ways we can we can break that dynamic down I say you know unblock as many people as you can tolerate is an important thing because you want to make the social media network more permeable to different ideas so it's not just about what you see is what other people see through you so you want to do that the thing you want to do is you want to engage in a human way with the content don't just read headlines and and re broadcast and share and like things read the content because the algorithms have a simple minded understanding of the content in our operating on this implemented understanding use your human mind don't be like an algorithm understand it content engage with any human way get to know particular riders get snow picket particular journalistic institutions and even if their institutions I would disagree with basically it's better for you to have trust through relating to human institution than it is to have trust simply by what did Facebook gives me you know what what algorithm search committee so says they hate more humanly online so that we can can counter some of the effects of the way algorithms make us more divided all right let's go to Thomas in la Hoya California Tom go ahead your check thank you for taking my call and Robert great information and tomorrow yeah I mean just wonderful questions Robert so one of the first applications of artificial intelligence was to M. dot nuclear command and control and the Pentagon you know decades ago realize that in order to have a nuclear exchange in thirty minutes they estimated that by nineteen eighty five that nuclear exchanges between countries would be thirty minutes long and you would have three billion that's B. as in Baker billion weapons interactions which means you had three billion decisions to make in those thirty minutes so right handed it over to supercomputers sand if Vance start official intelligence programs yes one general said it takes at least two seconds for general to make a decision then push a button that's way too slow when you're conducting a nuclear war at the speed of light and courses on intended consequences on that Peter Vincent Price who's been interviewed on coast to coast wrote a very interesting book twenty years ago documenting a number of nuclear incidences which were nearer lunches and the nuclear launch sick once spoken Russia and in the United States was initiated by computers okay and there was only one humans realized minutes maybe even seconds before silos were opened and missiles were launched there we were in a nuclear launch sequence so I was wondering if anything it's worse today then there was twenty years ago thirty years ago Janice wondering if you would comment on that it seems like this is what you're talking about there may be yet we need humans in the loop and we've taken humans out of the loop in terms of nuclear command and control and hello maybe we need humans in the loop the reason why we don't have humans in the loop yes because technically it's considered too slow if it I mean yeah it is interesting I I II I hear what you're saying I I I do agree with you and I think in some ways maybe is a little worse now because it's not just about the largest possible feel warfare nuclear warfare at the level of we now we have drives you know upon this vehicles that that basically a semi autonomous vehicles most of drugs are controlled by somebody flooded with the stick but they're becoming more autonomous in their armed with weapons and they go out and do actions you know far missiles that particular cars you know it's down to a fine level and there's real questions now about how much upon me do you give a weapon in basically making a decision in warfare and there are groups out there there's a a scientist and in here in the U. K. called little shark you will I know he's working very hard on this you were trying to establish international treaty about autonomy in weapons systems effectively that what's local closely called killer robots in this killer robot doctor occupy unique part of the military echelon from down at the drone level up to the the to the nuclear command level and I think it's very very important that people realize how simple algorithms are in many ways hell they can have a merchant affects how effectively you can program very simple like that because they interact with the complex world they have their own complexities they can basically make decisions you don't expect and therefore it's is absolutely essential I think in all important human interactions down from from from somebody get in their welfare check all the way up to nuclear command control to hand humans in the loop human interactions should be mediated by humans it's great to have great algorithms hoping that they should always be seen as helpers and helpers who were a bit simple minded and they need a real complex human being involved in this this effort to restrict the killed decision being made by an algorithm is very important it needs to be addressed at the international level with you and other bodies that NATO except for and people are working hard on that they're they're groups working hard on that and I think that that that their work is absolutely vital got Taylor now in Atlanta Georgia our first time caller Taylor go ahead Hey George Hey Robert what question all of the gentleman's question number the nineteen eighties we'll be working yes Sir do an act that was a kid accidentally not yet well you started or lines of command on accident okay woke up an old computer and started a nuclear Holocaust what so the union or not you think you think as of computers run ordinarily ninety out our programming yup and you think self Fatemi aware like programming or whatnot and lead to nuclear Holocaust countries either cyber command yeah I I think that a yeah the the great I love a on a science fiction movies I love them they're huge part of my life I see I think I've seen them all and I think that the great thing about science fiction is not necessarily that the science is accurate or will ever be accurate it's more that it causes us to ask questions about ourselves and I think a only in some ways it's kind of a field of philosophy it's still a fluffy this basically saying what is mechanical about thinking what isn't and and I talk about that a bit in the book in in the thing is is I think it with the war case scenario actually happen it's complicated on the one hand the idea of sentience and and kind of the robot being like a person as it is in science fiction and in war games that's really really really really far out in the future that's not real let's not really going to happen on the other hand could mistakes being made by algorithms that lead to catastrophic consequences yeah I think it's really possible and I think we're in a world where that it's become possible so I I do think it wouldn't be as cinematic as wargames it's it's something like that happen but it's not impossible that it could happen I I actually don't the the nuclear worry I have let's worry about then the kill robot scenario because I think that right nailed the debate is really about autonomy of small weapons like train tracks and things like that I think that's the the issue of the day but certainly one needs to be concerned about algorithms anywhere that make decisions that impact human lives in really vital ways what do you think of as a mas rules of robots those three and a great reason I love that stuff and and I think as most you know but but as I say about science fiction in some ways and think about science fiction being a basis for the next book the science fiction view of a on a kind of the future results you they all right being kind of a frame for the next book I think that sci fi is always about philosophical framing philosophical questions were that it is about prediction I think that a Ali is very far from being what we see in science fiction in general but I do think the exploration of the ideas is very very important and I'm glad that that all that so if I exist Gabriel's want us in Spokane Washington he gave go ahead I'm not saying this I purchased and mail and are all shook down on near the end and what I was looking one of separate phone records and rounds the life is not permanent well one yes yeah we'll do elaborate on algorithms is there something like a Grace Kelly algorithm that can influence and emerging American Indians socials monarch you okay I don't get that totally Robert how about you I don't think I got I don't think I really understood the question how far can and algorithms goal very far a in the sense that they are right already it looks think of it this way right now algorithms are deciding what needs you see.

Coming up next