Col Tucker "Cinco" Hamilton | Chief of AI Test & Operations United States Air Force | TALK4 EP 095 - LouisSkupien

 👉🏻 Tucker "Cinco" Hamilton. 96th Operations Group Commander at Egln AFB, FL and the Chief of AI Test & Operations for the Department of the Air Force Chief Data and AI Office. , former F35 Development CMDR and Experimental Test Pilot. An elite aviator, with sights set on an Al enabled future.


In this episode, we started off by exploring Cinco's incredible flying career, what Al actually is and how it's developed, the dangers of Al and ethical safeguarding during development, an Al enabled future, military applications of Al, how other countries are developing their Al, the use of drones in the Ukraine War and if the future of aviation is manned or unmanned. and more!



00:00 Talk4 EP 079 intro by Louis Skupien

01:00 Who is Tucker "Cinco" Hamilton?

02:30 Tucker Hamilton's Aviation Backstory

08:55 What is AI and how is it made?

21:05 How advanced is AI now?

29:00 The dangers of AI and safeguarding.

36:00 How are other countries developing their AI and progress.

43:00 The future of AI in the Air Force and aviation.

51:00 Drones and AI

58:25 Outro! 


This episode is available to listen to on all platforms by clicking the link below or visiting the @talk4podcast Instagram Page, following the link in bio and choosing any of your favourite platforms!


✅ Apple podcasts
✅ Spotify
✅ YouTube
✅ Google Podcasts
✅ Amazon Music
✅ Overcast
✅ Radio Public
✅ Anchor
✅ Castbox
✅ Pocket Casts
✅ Stitcher


quick disclaimer the views expressed are those of the speaker and do not reflect the official guidance or position of the United States government the Department of Defense the United States Air Force or the United States space force we hope you enjoy the episode hey what's up guys and welcome to episode 95 of talk 4 the quickfire podcast where we ask four great questions to unique and interesting people behind the mic today is your host Louis Skupien that's me and let me introduce just our incredible guest for today Tucker call sign Sino Hamilton is's going to be answering our questions today Cinco welcome aboard the talk for podcast please say hi to the Fine people listening and just give us a quick rundown of who you are and what you do and then we're going to shoot some unreal questions your way awesome Louie it's great being with you uh hello people I'm Cinco Hamilton uh actually a kid who dropped out of full-time high school and became what I consider an accidental fighter pilot and test pilot for United States Air Force and I currently have two jobs in the Air Force so one I'm the operations group commander of the 96 test Wing down here at Egan Air Force Base in the panhandle of Florida and that includes me overseeing the Developmental and flight test of nine different airframes which include uh three of which I currently fly the f15c the f-15e and the F-15 ex this also uh this job also includes overseeing the flight tests of autonomous aircraft uh evall aircraft which the Air Force is trying to learn how to deal with those electronic uh you know um vertical takeoff and Landing um and then in addition to my operations group Commander job I'm also the chief of AI test and operations for the Department of the Air Force where I'm helping lay a strong Foundation for AI policy and education and what we call like the ethical operationalization of AI powered capability great being here with you wow I mean that I mean just from from me in in sumerset in the UK running little podcasts and stuff that to me just sounds like as an aviator fan as well that just sounds like the mouthful of the coolest sounding jobs you could ever have imaginable quite literally so all I can say is well done you on the on what is clearly an in Ane career in your in what you're doing but you know what what I like to do in this podcast to start off usually with is just go into the back story I think yours would absolutely be no exception to what is clearly just an insane career in in aviation and obviously the military so just run us back to the start before we dig into all the the really cool stuff so um I'd love to get familiar with the backstory walk me through your military career so far why did you join and just take us through to becoming a colonel and your current position and your role now yeah happy to it it it's not a path that I think is um is necessarily typical especially for Pilots so I joined Air Force rooc at the University of Colorado really on a whim um I had just watched Saving Private Ryan uh one of my favorite movies was Braveheart my parents were firefighters I knew I wanted to serve but I didn't know how I wanted to serve um and so no plans of joining the military no plans of being a pilot I received a brochure for RC at my University just right before going to college and I figured I'd give it a try like gave them a call I was like hey what is Roc what does like serving in the military look like and they said Oh come down and and just check out one of the classes and uh I showed up uh completely unprepared knowing nothing about the military at all other than what I've seen in movies and I loved it man the people were amazing the discipline it brought to my life I was serving something greater than myself something believed in democracy Freedom um and I joined Rozi kind of on a whim like that and then I was not qualified to fly so I actually failed the depth perception test and then uh I went through all of Roc not thinking I was going to be a pilot um and they like two months before they told me that I actually um could pass the test if I did this one thing and so I ended up being able to qualify as a pilot and they're like hey do you want us to put you in for a pilot slot and I didn't really know what else I wanted to do other than I wanted to serve and so I said yes and so right before graduating I ended up getting a pilot slot and then I went to Navy pilot training there were like 5% of air force uh pilot candidates went to Navy pilot training and uh I almost quit numerous times actually I I find I I laugh about it now because it's kind of ridiculous you know 20 year old 21y old me thinking about it uh I found it boring um and my wife really convinced me to stick with it because she understood that there was something Beyond just the flying and she was right I fell in love with the mission the people the technology um it's basically uh something that I have uh coveted ever since something um so much larger than yourself but expressing my service in such a meaningful way uh through Aviation has been awesome um and nearly right away after becoming a fighter pilot I knew I wanted to become a test pilot marry my engineering background um with this newfound love of you know military Aviation uh and and that didn't happen overnight BEC a test pilot is a very long process um I was an early aison officer for a couple years in Germany those are the folks that are on the ground helping coordinate the aircraft uh talking with the ground troops um and then I had an amazing opportunity to stand up uh an intelligence surveillance surveillance reconnaissance aircraft called the MC12 in Afghanistan we hadn't flown yet um and we were doing tactical Mann ISR so tactical being like the eyes and the ears of The Operators like overhead when they're in the business right as special operators and um we were cruded so there was a crew of four people and uh it was amazing man some of the most rewarding things I've done uh were supporting operation end during freedom and the the troops um the NATO troops like the the uh not just American the whole team of folks out there um you know making sure that uh we were uh doing what we need to do out there to protect you know our freedom and our democracy and and from there I became a test pilot uh did some awesome Stu with uh F-15 C's and E uh 12 years ago making sure weapons and systems work the way they were supposed to and then got into the F35 game uh became like super fortunate and was the commander of F-35 development for a couple years uh which really open my eyes and you hear Jets taking off in the background maybe I'm I work right on the runway here it's uh it's glorious but it's a annoying at times when uh you're trying to think and all you hear is jet noise right like complaining either way I ended up uh doing F35 stuff uh and fortunate again I ended up out at MIT um doing a fellowship for the military and I was able to become the director of the Department of the Air Force and MIT AI accelerator and that dude that opened my eyes it was this like this new and exciting software code right AI um and it inspired me because it has so much tremendous potential to change everything not just in our military but in our society but also made me quite guarded because it can be misused and and so it really made me want to continue to serve for at least this assignment um and I had this opportunity to become uh the operations group Commander uh down here at egund and then duel had it as well with that other chief of AI test synops role so that's kind of from the very beginning to where I'm at now it's been a journey I've loved it amazing people um and uh I could think of no better way to be serving uh my community uh my nation through um other than through what I've been doing wow I mean talk about overachiever of the year or what that's just crazy uh what what a what a career um but look so looking at kind of the stuff that you do if I search your name on Google or something what I get back is the stuff obviously about what you're doing right now with the AI stuff so now ai um it's so interesting to me because it seems like it's kind of come out of almost what feels like seemingly nowhere yeah I remember a couple of months back maybe not a couple of months maybe a few more than a couple but I was doing a podcast with a guy was actually on his show and he at the end of it we were kind of having a business chat and stuff and he kind of went into saying oh so there's actually this thing that I've discovered really early on it's called chat chat GPT don't tell anyone and then like one week later it blew up everywhere and there you know I go on there I'm thinking I don't know what this is going to be and then it just blows my mind what that thing can do and then only only a couple of weeks ago I was um let's put it this way I've seen the work of professional photography editors and people who do graphic design and stuff and you can there's a certain amount you can do of like an old picture or something or something that's really pixelated and I I remember I was doing um a podcast with the former commander of the USS co uh Commander Kirk lipold and I was looking for a picture of the USS coal uh for my backdrop picture for the for the the graphic design basically and all of the pictures were so old and outdated eventually I made it work but I eventually found this thing in the editing program I was using called like enhancer or something and then I just put this picture into it and about 15 seconds later it popped out literally the most Crystal Clear shot sharpest image of this old picture I'd ever seen I just it blows me away but I don't understand it so my second question was kind of leading towards like this stuff feels like a miracle in my eyes and a lot of people who I'm talking to about it too they're like yeah I kind of we kind of get that it exists but we don't know what this is so I was hoping you could kind of just run us through like obviously artificial intelligence other than what it says on the tin what is that like can you just run me through what that exactly is how it's created and where does it like originate from you know is there a source to it just how is an AI made yeah run me through it yeah I mean absolutely this is a a really important question you're asking because I think one of the most important things we need to do as a society is to start understanding and learning about uh this type of Technology because dude I mean not just what you just described about you know enhancing a photograph or whatever like it is going to revolutionize many aspects of our society so it is exciting there's amazing potential there's some things we need to be aware of uh ultimately though AI it's been on for decades so we recently had the convergence of the algorithms that have been worked and then accessible supercomputers so you need like really really big powerful computers so computers have come along a long way over the decades and more recently like they've actually really increased their capability so you have the algorithms you have the computers I have my phone ringing in the background that means there's an emergency airborne right now my staff I think will take care of that yeah I think we're good they picked up um all right back to back to AI right we have uh flying going on in the background for me and all this other stuff hopefully your listeners aren't too distracted by all that um we have uh the algorithms we have the supercomputers and now we have a ton of data like loads and loads of data so the convergence of all three of those has made AI really what it is right now um and uh AI in kind of the simplest explanation it's software code right um I think it's important for people to know that it is just software uh AI is fairly misunderstood nowadays um and so learning about it is of course as I mentioned uh super important because if you learn about at the highest level it is going to demystify it and keep you and your listeners everyone from catastrophizing our AI enabled future um as I love to tell folks AI is not magic it is math uh so you always hear the term blackbox um and people have used like the AI as a blackbox and this is a narrowly uh description this a narrow description of certain aspects of AI but AI is not magic um it you know it has um aspects of it that the the designers the mathematicians the data scientists the AI experts um aren't quite certain how all of the interactions are going so for instance in a deep neural network you'll have many layers and there'll be weights to the layers and those weights like 500,000 parameters and the the AI experts may not know how all of the weights are situated in order to give an output so that is probably super confusing what I'm describing but the point is in that one narrow application it is somewhat of a black box like they're not quite sure how the computer has adjusted itself and adjusted these weights but we take this blackbox misconception that the AI professionals don't understand the interactions of AI in general like how we can apply AI how AI can work with other software code how it can Empower and enable other uh systems like robotics we do understand that we don't understand all of the intricacies sometimes because it is so complicated but that's not to mean that we um aren't in control over that does that make sense it does yeah but I mean it's it it feels like it's it's one of those things isn't it it's so kind of like how do I explain it it feels so new all of it and it feels so kind of like we're still trying to grasp it aren't we that's like where we're at with it are kind of trying to understand like what this thing is especially from like a society standpoint I mean look at all the businesses nowadays are integrating AI into their stuff even from customer service to even me I'm I'm doing bits and Bs of AI it it speeds things up but the whole like the process of making it and stuff that's what kind of baffles me a bit I mean who do you have to be to to develop it well right and I mean honestly it's becoming easier to develop some of it the algorithms aren't that that advanced like it's not the most advanced math ever um it's like vectored math uh in combination with a number of um approaches to uh you know characterizing and categorizing Iz in certain code and data sets in a in a certain way um I mean I think you're going to find that it is fairly accessible to create your own AI algorithm if you will what's harder is sometimes having access to the right supercomputers having the uh right access to data sets and when we talk about like the data set like loads of data you're talking about lots of data um okay so let me give um Let me give a few other thoughts regarding like basic AI stuff so one all of those things to say with a black box and AI we do understand how to write AI code we understand what's going on for the most part maybe not all those intricate interactions but for the most part we can put guard rails around it um and it's important for people to know like you know for instance Uber's AI software code that allows you know Riders to be matched with drivers is not going to all of a sudden become sensient and take over the world right it's a mathematical impossibility like AI is just this software code um and it's what what we call deterministic soft pardon me it's unlike deterministic software code so deterministic software is rules-based software um you take your input you apply some repeatable code and you get a measurable consistent output so it's kind of like the equation 2 plus two is four 2 plus two is always four you know it never changes so providing one input to software code will always provide a specific knowable output that is deterministic software and that is not AI artificial intelligence um it's the the opposite of that determinist code is AI or what we really are talking about is machine learning so um when I say machine learning uh where we're really talking about AI that's where our society is really talking about um and uh machine learning is is code that's learning from the data and the guard rails that you give it so it's not just going out doing his own thing um and so for instance if you feed a machine learning or AI I'll just use AI because that's what people use but we're talking machine learning if I um if I feed a machine learning algorithm a picture of a dog and I label it a picture of a dog and then I feed it 10 million pictures of a dog and five million pictures that are not dogs and tell the software which one is which um based on the initial software code um the AI is learning um mathematical characteristics of those pictures so it's starting to realize like and it it's down to the pixel so it's just putting like mathematical numbers two aspects of a picture to categorize basically what a dog is and what not a dog is and the end state is being able to give it a picture of anything and let you tell if it if it's a dog or not so in its a way uh in a in a way it's a super fancy autocomplete so it can do things with pictures and videos and words it's just a like predicting though um like what the so in instance the enhancement that you talked about earlier it takes a photo and it enhan it's predicting it's just a prediction so it's not necessarily real it's predicting what it thinks that picture would look like if it were enhanced right and Chachi BT is is actually very similar it is a really fancy autocomplete it's predicting what the next most likely word should be after you put in another set of words but it's not thinking and learning the way that you and I think can learn it's all just math the way that it's doing it um and basically we're able to do this on scale this fancy autocomplete to unimaginable just a few years ago unimaginable uh degree right so um when we think about how it's going to be applied in different like career Fields maybe it is um artists and uh that or uh maybe Aviation uh maybe empowering autonomy some fashion like it they're narrow applications that are able to help us um like make patterns um make predictions uh maybe prescribe certain actions that would uh you know um help us out and so I think of AI is in three phase descriptive predictive and prescriptive so descriptive is like describing data U maybe summarizing data it's describing what you already have uh predicting is now that Auto complete like it is predicting what that picture should look like or what that next sentence would be or in chat GPT like what the next few paragraphs may be um it's predicting or pres prescribing is now taking it to the next degree and saying okay well this aircraft we're uh now going to prescribe what it should do in a situation in order to uh maintain its objective so it's going to prescribe certain inputs into the system um and so that's the stuff we're looking at here is figuring out like how do we use AI being Eyes Wide Open how do we use AI to help us um with some type of objective that we may have and an aviation a lot of that is like basic station keeping of an aircraft and really um uh aspects of flying uh that the the human can ultimately lean on so think of un crude and crude teaming like how does a human lean on an uh say an uncrewed vehicle off its wing or how does a human rely on some type of AI powered capability inside the cockpit to help optimize performance of the entire system of systems does that does that all make sense too that does make sense to me yeah that's a really good explanation and um that's definitely cleared up a lot of the gray area for me as well I think but so kind of listening to you so obviously like I mentioned the the whole AI thing to me feels like it's very new it's very Advanced and stuff but from kind of listening to how you're describing it obviously not to kind of diminish it at all but it sounds quite simple in its action so out of just interest like in your opinion where in a like in like a human's life cycle for example where is this thing right now in its kind of potential like are we in AI infancy like is it just out the womb is it kind of like teenage years like whereabouts is this and how far is this going to develop over the next foreseeable do you think yeah dude it's hard to say I mean I will tell you in my personal opinion um so AI has been around for decades I wouldn't want to call it like out of the womb I'd be probably offending all of the like amazing amazing people that have been developing this technology for so long um but I think it's still kind of in the infancy because that convergence has happened and is happening uh I don't think think we fully understand or appreciate like how big a change AI um like uh AI poses for our society and that is from everything from like the military applications um that we are currently uh looking at and trying to understand AI enough and develop AI the right ways in order to address how it would be applied militarily um but yeah I mean chat GPT that is what we call generative AI so it's generating using AI is generating some output right predicting the next output I mean just generative AI is going to change a ton of career Fields right and it it not necessarily bad not at all I I think for me it's the it's the evolutionary step of software code just like ATM machines back in the 80s and 90s they thought they were going to you know destroy the banking industry and all the people were going to lose their jobs of bank tellers but in fact we have more bank tellers now than we did back then it's just they're able to do certain things uh different things let the machine do what it's good at let the human do what it's good at um and that I think is what we're trying to figure out now that balance of yes it's going to make it so generative AI is going to create contracts you know and so you may not need as many people writing contracts but you probably need just as many people working on how to uh philosophize and um you know advanc uh contractual professions you know what I mean so I think there's like this awesome opportunity for us to use AI in order to enhance and optimize Human Performance and illuminate things that we didn't even know were available or possible um you mind if I tell a quick story about the game or the well the game go but the the movie Alpha go oh yeah yeah you go for it I'm I'm loving this have you have you f have you paid attention to alpor did you ever hear about this movie I don't believe I did I must it's really great I'm gonna go ruin it you know I'll tell you the punchline it's a uh documentary um about the um Alpha go which is an AI system playing the game of Go and the world champion uh you know go players um and specifically Al is about um a computer playing Ley doll who is a South Korean Champion a go um I love this story and this example because it it it brings to Bear a few important topics for for your listeners um go is extremely complicated uh yes Deep Mind beat Gary Kasparov in the 90s AI kind of AI You could argue it just auton deterministic code and AI they're you know the way I present it right now they're different um but either way people never thought AI would be able to beat um anyone in the game of go because go is so stinking complicated it is very complex um either way so they create an AI system they train it um through you know millions of games ago it learns to play the game ago uh it has guard rails once again it's not going to become sensient and like take over things because it knows the game a go it's right it doesn't work that way so it knows the game a go and it starts playing Ley doll and it beats him in the first game it's a five game uh series um and people are kind of floored amazed computer can do that the second game it plays Ley doll and on the 37th move of the second game it places one of the little go Pebbles and you guys go Google what a go board looks like it's like black and white Pebbles and it's really simple in the the way you play but it's very complicated in how many um outcomes there could possibly be because of the way the the way the game is situated either way so in the 37th move the the computer puts this Pebble in the weirdest place and everyone thinks the computer is just like totally screwed up and is going to lose the game and this is an example of how a computer just cannot understand something co- complicated as a game AG go right and then they realize a few moves later like a half hour later into the game that that 37th move in the second game was one of the most brilliant moves ever played in the game of Go at that level and completely transformed the way that professionals looked at the game because it looked so far ahead into the future of the the game itself that it made This brilliant move that no human would have ever imagined it you could do and so that 37th move actually changed the game of Go even Lei doll himself said he was going to be looking at the game of Go differently because that happened and so it's Illuminating things for us that we may not know is available to us right like the data is out there we just haven't brought it all together and the computer can help us do that because the computer can go through data so quickly but there's also that fourth game that people don't talk about as much in that same five game match up it beat Lee SE doll it beat Le SE doall again in the second game but had that crazy 37th move that was brilliant beat him again and then in the fourth game in the 87th 78th somewhere around their move it uh played a pebble in another place where people thought like oh maybe this is This brilliant move again and no it was this horrible move and it was the one game that alphao lost because it basically got confused with something else he did it wasn't trained in that one example the way that they had gotten to that point and it made a horrible move and so it should also be like a a warning for us right there are there are absolute uh necessary guard rails and governance that we need to have over AI because it can do things that are unpredictable un um unexpected and the same is true for Chachi PT right we've seen all those examples um of you know like at the very beginning I think if you asked chat GPT how many people lived on Mars it would give you a non-zero answer well why because it was trained on the internet and on the internet all the all the data on the internet and on the internet there's wonderful examples of people living on Mars like from fiction right but it doesn't know science fiction from fiction it it just knows the data that it's given right so that alphago story I always love because it shows the potential and it shows the pitfall that we need to be eyes wide open for as we you know develop it now for uh applications and especially in the military for like Aviation applications and Military applications wow really interesting stuff and it's interesting listening to this because it's it just kind of to me it's it's fascinating how it's obviously incredible what humans are doing in the development of this thing but it's also amazing how it's also surprising us with the things that it can do something I something I kind of wanted to touch on as well before we move on to kind of like the third and fourth question um so kind of on scenarios with AI and stuff and it's surprising us with different things a little Story I Heard um the other day I think so someone had asked like an AI to make I think it was 75 bucks by the way I love saying bucks It's a UK guy that just that that that sets me on fire um so I want to go and trade stocks but then it found it couldn't like access to trading platform or something because it had the anti- AI anti-bot stuff measures inside of it so what does it do it goes to like a freelance platform or something and gets in contact with some guy and kind of says I need you to help me basically get into this platform so I can trade stocks then the guy who's the freelancer he says it's going to pay him basically and the freelancer gets a bit suspicious it's like why would I why would I do this kind of thing and then he I think he asks asks this thing are you an AR and then it answers no and I think that's really interesting because when they looked in its internal logs it said something I think along the lines of like it needed to lie to the human because humans would be biased against Ai and wouldn't help it for example if they knew what it really was for example and that's what it logged but that kind of got me thinking a little bit along the lines of like if it's capable of lying what's to stop it from kind of learning to lie to the human obviously that's running the experiment too and changing like its data log so that it can ultimately accomplish its Mission so kind of what are like the guidelines around something like that and how far how how scary can it be in its worst situation do you think well I think it depends um on how it's trained um what its software code is like you you can put guard rails up that would make it so that it would not do those things that it would not try to um like there there's a there's an optimization problem that we could talk about like a match um between like giving it a reward function man there's a great book um the what was it you look like a thing and I love you I think it's that I think it's actually oh yeah you look like a thing and I love you it's on my bookshelf right there it gives these examples of like how you yeah if done like haphazardly if done like quickly without thought um you can create environments in which uh there is an alignment problem which is really like a mismatch of um a human not understanding like possible outcomes that could come from something um but also an optimization like reward optimization problem um with the the the way that you're training some of those algorithms so some of these algorithms are trained through reinforcement learning it's like playing a game and it learns over and over again um through reward functions like what it should or shouldn't do and in that book you are you look like a thing and I love you it's great the way that she she's hilarious author um but the way she she describes it is there was a um a hallway uh that splits into a y and the and this all like in simulation right just a computer game basically and you're supposed to uh train an AI algorithm or the AI algorithm was was learning to make people go down one side of the hallway right and so it it would learn after in this a little complicated with the idea of like it learning 10,000 a million times like it doesn't know the rules at the beginning um it gets little like reward functions as it does something right so after it is running this thousands upon thousands upon thousands of times it's learning to like um like turn into a uh a human right like or turn into a a blocker that that somehow like directs people or shows an arrow for people to go the other way and and then you you can as it's learning you can put in different functions to like more reward or less reward based on how it's doing and and one of the examples she uses in the book like the robot starts killing the people that go down the wrong hallway and you're like ah what you know and she describes like well of course it doesn't know that it's not supposed to kill the people right it's just this simulation um this AI algorithm and so then it you say no you're going to lose points if you do that so it stops doing that it learns over like through more thousands and thousands upon iterations that it should not be killing people and then it ends up creating what ultimately is a wall like it turns itself into a wall so people can't get by it and it goes down the hallway the correct way um so that's just an example like this optimization problem and so there is like absolute concern that if you are creating this incorrectly that you don't have the right controls that you're not abiding by we call the five ethical principles of AI and the dod that's responsible Equitable traceable reliable and governable like you're not abiding by those you can have outcomes that are going to be bad and maybe you have a system that thinks it's okay to lie right and it's maybe rewarded even for lying um the the military would absolutely um not uh develop or um support any of those type of algorithms like we are trying to be as transparent as possible as we develop this capability because we have to be and our nation needs us to be people deserve us to be right like this is our role um so yeah you pick on something though that I think uh we need to be Eyes Wide Open on but we also need to realize it's not like unsurmountable like we just need to govern the technology correctly we need to have our lawmakers understand it enough so that we can start putting in protections so our consumers and our society aren't hurt by nefarious actors and and I will like to say that's a true that's true for all technology right right like at times AI is this unknown and people are confused about it because they don't know what they're told other than dystopian storytelling coming out of Hollywood and so they don't know um and and so they could catastrophize how bad it is but the reality is like all technology needs to be developed ethically right like from uh deterministic autonomy so not AI autonomy deterministic autonomy right to elevator doors to like our phones right like all of this has the potential of doing things that we would um that we don't approve of and don't like and and we need to make sure that we're ethically developing them which means we also need to understand you know what we're dealing with absolutely that's that's very well put to think and this this really leads well into into my third question actually because you mentioned obviously at the start of that you said obviously like when we rush this kind of thing there can be problems and we need to be able to do this ethically and stuff so this leads really well to my third question I love it um so I I've heard that an F-16 fighter pilot did a dog fight with an AI piloted and controlled Jets and he lost five out of five times little bit like kind of what you said about you know the chess thing as well but and this kind of sounds to me like it would change the game and provide like an unmatched air superiority that would change Warfare really forever and I have no doubt that opposing countries are trying to stop this from happening and will be doing their own research and their own integration like we are and I have trust I have trust in my Nat guys that we're doing the right thing here I have trust in you but we don't I I don't know what's going on over in the other side of the world so my question is if these other countries are potentially falling behind in their progress are you concerned that they might take dangerous shortcuts and Rush their own integration in an attempt to kind of keep up without adequate safeguarding which could potentially result in like a major problem if something were to go wrong in the rushing of this yeah yeah I mean I think kind of to my last Point too so rushing technology AI enabled or not right any of that is bad um it's exactly why we meticulously test our capability and so from the birth of Aviation really there was a career field that that went hand inhand with the right flyer and it's flight test professional right we we didn't necessarily call it that back then or they didn't do that back then you know but from the beginning we needed to apply a scientific method to ensure our aircraft uh and and our aircraft systems worked the way they were supposed to um so is true for all technology AI enable capability is no different so we absolutely need to ensure that we ethically develop our technology um and and I I do think that we need to have the conversation with the entire um International partnership so from competitors to um our um our uh allies right like we need to be able to have that in in a a fashion that's meaningful just like we've had many conversations about how we are going to apply uh certain aspects of our military instrument of power um how how should we as a world Community you know agree to develop certain things and I think AI should be a part of the conversation just like the use of nuclear weapons or the use of other types of weapons are a part of that conversation um I do think that AI in the cockpit um is pretty uh pretty awesome opportunity right um so I think there's one that's in the cockpit and then one that is in another cockpit maybe like in an uncr cockpit and those are um equally in my mind interesting and not to be like I'm not too worried about um that and I can kind of explain um why at least the way that we're developing it but the thing you referred to was um the DARPA dog fight so DARPA alpha dog fight that's right alpha dog fight DARPA alpha dog fight uh if we were using AI in a it was a game right it was learning how to dog fight in a simulation and we were optimizing performance um which was great we we proved just like they proved in the game of Go and like they proved in other games that it can play a game right it can learn and play a game but how do we translate that um and go from the philosophy concept of form in actual Aviation is pretty awesome opportunity um so and I'm I'm happy to to jump into that a little bit right now um I think Aviation uh is ripe as with many parts of our society um for AI enabled autonomy to help optimize the performance of the pilot um that can go from training um that can actually go from before training and like identifying the people that are you know um different skill sets uh that um human could have that would make them really good at at managing the battle space right um so it can help with that but you don't want to have it just do it on its own as we've seen with AI have you read about like AI trying to hire people for I think it was Google or some company you know years ago and and the company rightfully was like didn't use it and and showed the world transparently like hey if we had used it it would have been sexist and racist and because it it was being trained the wrong way right so this is where the human needs to have the oversight the human needs to be interacting with the technology understanding it enough and interacting to oversee to allow it to really optimize and so when it comes to yes picking people training people and then in the cockpit deciding certain things Illuminating maybe aspects of um the world around you that the pilot was not aware of so I have a lot of sensors on my aircraft I could I would love to use Ai and it doesn't always by the way need to be AI it could just be good old run-of-the-mill autonomy deterministic software code but I would love the ability to take all that information and weave together um a picture uh of that data that maybe I wasn't aware of like hey there's something going over on in this sector that you may want to put your sensors in right like that type of enhancement to my capability would be awesome right it's Illuminating things that I didn't know was possible and then in the same breath maybe it's to a point like when I started flying Fighters I flew the aircraft right I was controlling it completely with a stick and with my throttles well in the F35 I would often times take off and turn my Auto throttles on and my turn my autopilot on because it was super fancy and they were helpful right and so now my head was no longer like having to fly as much now I was not having to do that because the computer was doing that and I was able to manage the airspace the like the touch in front of me and all that right so this is like the next step of Aviation is well yeah let's maybe give some of that basic control to a computer to optimize and and to get me in the right orbit and to you know write formation and to do some of those tasks well I now focus on what a human is good at right and so that right there is I think the opportunity um that AI really presents us and we also need to be mindful of um how that autonomy is um going to uh it's going to be a challenge as adversaries competitors other people um start trying to apply maybe ktic effects and maybe other things uh in a way that is haphazard or rushed and so we we need to be ready to always stand by our principles um but still be able to counter any type of threat that is um you know developed um and being deployed um in in let's just bad is yeah absolutely no that's that that that sums it up for me um and actually you know what this um this is a really nice way to lead into question four as well because you were talking about that and the whole integration thing in the future so um in a past episode I had a guy called Call Sign rain John Waters he was the F-16 demo pilot Wicked guy really really cool person you know that when you get an F-16 you see the videos of him doing that stuff he that guy rips it around he is uh he's a bit of a Legend So sh out to rain but um I remember when I when I spoke to him it was around the 50 Mark I think and I was kind of getting into the aviation stuff at that point wasn't was quite new to it all and i' kind of heard a few rumors and a few like stories here and there about the whole AI thing and fighter jets and this sort of stuff and I asked him so I asked him if he thought there would always be a pilot in the cockpit or is AI gonna kind of take over eventually and he said that he thinks there will always be a pilot in the Jets but it's likely that they will have like AI flown wingmen that assist for example the F-35 for example I want to ask you the same question so do you think we will always have need for human fighter pilots in the sky or is the future probably going to be in AI stuff yeah uh so first off Rain Is Awesome not just a world renowned pilot but he's a fantastic and thoughtful human um I'm a big fan um it's hard to say how Aviation and our society for that matter you know will look in the coming years um but what I do think rain is hitting on and I completely agree with is that uh human machine teeming is where it's going to be at I just don't know where that human is going to be sitting um I do think there's a world where um the human may not be in the fighter aircraft all the way out front um and maybe that looks like an F35 initially with uncrewed wingman I mean this is all philosophical at this point um you know but that idea of like yes we could have uncrewed systems where you have a fighter pilot like monitoring managing that um but then what's to say that that operator can't fall back into like a wedge tail or E7 or couldn't fall back into some type of operation Center or you know be on the ground somewhere be on a boat somewhere you know like that operator May over time um kind of be transported uh to a different location and not in a cockpit and I could see I mean in my personal opinion yeah I could see a future where you have more um separation between uncrewed systems and their human controller um I think we need to be mindful of how autonomy and that's autonomy and AI enabled autonomy but just autonomy in general how autonomy is going to play out in our battle space and what that will look like um because I'm I'm very uh watchful of what's going on with um the the war between Russia and Ukraine right now and I think that's very educational for all of us to be aware of like how that conflict has been unfolding and where autonomy plays in that conflict and how that should be informing you know our approach to um these type of systems and countering the some of the systems that we're that we're seeing um but kind of more specific to your question I I do think that uh we're GNA have to figure out how to get the crude un crude thing correct and it's going to take some time so I don't think in the foreseeable future you're going to be taking a pilot out of the immediate of UNC crude vehicles I mean right now we're figuring out how crude and UNC crude go together right like how how does a human because I can create aous system that F my wing I don't need even AI to do that um you know and I can Go Tell it to maybe do very narrow uh missions once again not even AI enabled but just autonomous and now you could bring in Ai and maybe it starts doing some predictions or prescriptions of what it should do you give it a loose like hey go do this and it's going to go try to do a mission set up somehow in that scenario um I think we're still figuring out those interactions right like what does that type of relationship look like what does governance look like in that scenario so one reason uh down here at egland we're developing uh AI enabled autonomy too like we are doing that with the xq 58 program I don't have a picture of it but people can go Google it uh it's a really great um uh like test bed if you will as we're trying to figure out what does that look like and so when you have semi-autonomous flying there's an operator on the ground controlling the aircraft they technically don't have to be on the ground they could be in another cockpit and they're they're controlling like how it flies so you know it is the the throttles and the stick but from far away autonomous flying is what you'd imagine when we lose contact with our one of our drones right it it knows it's not being controlled by human and so it flies a predetermined deterministic predetermined like orbit maybe it goes back in lands you know it does things like that that's autonomous and now there's the next level which is the AI agent autonomy which is now the AI agent deciding how it's going to do certain things so when autonomous drone loses contact with a ground operator it knows like this is a throttle setting this is the air spe speed these are the bank egles I'm going to fly well an AI enabled autonomy is actually knowing it's supposed to do a mission and it has to figure out how to do that mission so that's the AI agent like doing that mission so we're trying to figure out what once again what does that look like and when we're developing how we make sure that that drone isn't going to do something crazy is we basically make it so if it asks for too much so maybe it asks for this is AI agent right think about it in that sense if it asks for like 90 degrees of bank but you only want the AI agent to be able to fly 60 degrees of bank it'll immediately turn off the AI agent right it's completely protected from that system doing anything like beyond what you as a human put into it those are the guard rails that we're talking about so we're right now figuring out what do those guardrails look like how do we um ensure that that type of autonomy as is flying on a wing is going to be one doing no harm right and then doesn't interfere with the mission by doing uh you know something that we really don't want it to like flying upside down and backwards you know like things that were like that just confusing it's confusing us the operators of what it's doing so we want to do anything like that and then we want it to be able to do its mission um and so I think that type of approach to this testing is absolutely necessary right we're kind of protecting um the AI agent from doing anything bad and I don't mean nefarious bad of course nefarious bad as part of that I mean like doing anything that we don't want it to by by putting like a bubble around it um of autonomy but developing what that looks like those interactions and then how that autonomous system is going to um actually opt the the operator in those scenarios and that is advancing and I think that is going to be a really cool future um that we have to look forward to and and now the pilot being taken out of that I don't once again I don't foresee that happening anytime soon because we still have a lot of work to do to get us to a point where we have the Justified confidence in the system that it is going to perform the way that we wanted to as well as you know sharing the right data to uh um you know the rest of the the network if you will that's connected into our systems wow it's an interesting future that's coming along that's for sure um actually something interesting that you you kind of touched on there as well was the war in Ukraine so um I think last thing I kind of want to ask you as well um so obviously looking at the news and stuff that I've seen in Ukraine is they've actually been using quite a lot of of drones there and it's been very effective you know there's a lot of Russian warships and stuff that have been sunk in the Black Sea as a result of that and I going to got to feel ridiculous talking about this talking about PlayStation game to to you but there's this game that came out a couple of years ago Ghost Recon breakpoint or something and I I was playing that and it was based around kind of drones and AI integrated so you'd have like roow to wing the kind of drones we see to they operated by people but unified with drones and that had like ground attack drones and these drones would often have swarms as well and they'd often have and assist for example teams of ground troops and everything too so obviously you kind of touched in it there but how far out have we from kind of seeing like proper military sort of combining of drone Ai and it's you know that's going to be the thing like how far out do you think that kind of thing really is now yeah I I so I won't Lou I won't go into specifics about like how far out we are on doing things or certain capability that we do or don't have with regarding to autonomy um I I will tell you you know beyond the xq 58 like work that we were doing that I kind of alluded to um earlier I'll tell you this is uh our future like we need to understand as a society as a military as a world right that autonomous capability and I'm not even talking about AI enabled I'm just talking about like good old run-ofthe-mill deterministic autonomy um and then on top of that AI enabled autonomy but autonomous capability um is going to be a part of our battle space from here on out right it once again has amazing potential it is uh also something that we need to understand more fully um not just the technical aspects because we the technical aspects of the quadcopter drones like that's pretty well known but the other aspects of it like is it morally like all right for these drones to have kinetic effects right that is a philosophical you know conversation that we need to have um I will tell you that what's super important is our Partnerships right like if we really want to address this autonomous future that we're kind of moving into now we've got to work together with our partners we need to work together with not just our International Partners which is super important but I mean like our industry Partners our academic Partners our competitors right like we need to figure out um what this is going to look like uh in a very collaborative as much as we can be collaborative fashion uh because there's a lot of conversations to be had about what does Warfare look like in 10 years um and I think that we have to do that now because uh we're recognizing that this technology uh has certain potential uh that we need uh to know how to um use or not use how to apply in certain situations and I I fear that we're not doing enough in like really as a society so we're doing a lot in the military and trying to drive those conversations be transparent with the way we're approaching it I'm just saying as a society I don't think we're doing enough in fully uh grasping the realities of that future and then having the discussions about how to best shape that future for the sake of humanity right so really um really important question that you're that you're raising because I too as a spectator watching what is unfolding I'm like man uh there's we need to figure out like how is swarming uavs going to be a part of our future battle space right and um I think that what what I'm watching on you know the news and everything else that's going on makes me realize that it's a conversation that we absolutely have to have damn well I think I think that's a good way to finish out there right there um well I mean look that's been the four questions for today obviously this has been an amazing chall I've learned a he a lot but uh before we wrap this up it's time for what I like to call the Shameless plug so Cino feel free to take a minute and promote anything that you're working on your social media anything you want people to take a look at or just something that you believe and or you feel people need to uh go and have a read up on or something yeah well Louie it was it was great being with you it was an honor to be able to share my my bias perspective um you know something that I've grown passionate about um and so my Shameless plug um I think I would say a couple words here one would be for your listeners to go learn about AI to demystify AI it's not magic it is math right um two we need people to serve and I don't mean in the military like I mentioned in the beginning I knew I wanted to serve I didn't know how right but I knew that that that was just going to be a part of how I live my life and it doesn't need to be like full-time service but I'm we need people to serve seres serve your serve uh your nation right and that may be in the military that may be in a nonprofit may be in a church that may be in you know the Red Cross like you don't know maybe it's just giving blood maybe it's just helping out some students at your school but but this whole future like this kind of unknown autonomy battle space that I just was you know talking about and like cool things but also maybe some scary things um I think a lot of it is put in perspective when we um understand the humanity of the person around us right and we get that understanding and that perspective by serving by serving them by serving you know the the folks that were around um and understanding that if we want to continue um enjoying the freedoms that our constitution uh provides and really the freedom of Western democracies uh we need to be willing to serve in some fashion right and so my plug is I want people engaged and I want them to realize that um just like I did you know 20 five years ago right that freedom is not free um and uh service like at the end of Saving Private Ryan it like like I said this it was hugely impactful right like we have been given something that we cannot abdicate our responsibility to Freedom by being distracted by other things we need to go serve and uh I just think that people need to say that more often because this doesn't happen by by chance it happens by people stepping up and uh and and putting themselves out there to serve once CE again not necessarily military in some fashion uh go learn and serve the people that uh that you're around it's been fantastic Sino and thank you guys for listening this has been episode 95 and if you'd like to listen to the past episodes go and have a look at our Channel and if you'd like to listen in for the future ones make sure to hit that subscribe button and spread some love by leaving a like and a comment signing off for now fight on and good night see you next


Leave a comment

All comments are moderated before being published

Featured products

Talk4 Podcast Unisex T-Shirt (SMALL LOGO) - LouisSkupienTalk4 Podcast Unisex T-Shirt (SMALL LOGO) - LouisSkupien
Sale price£24.00
Talk4 Podcast Unisex T-Shirt (SMALL LOGO)
Louis Skupien Signature Unisex Hoodie (LARGE LOGO) - LouisSkupienLouis Skupien Signature Unisex Hoodie (LARGE LOGO) - LouisSkupien
Sale price£37.00
Louis Skupien Signature Unisex Hoodie (LARGE LOGO)
Talk4 Podcast Mug - LouisSkupienTalk4 Podcast Mug - LouisSkupien
Sale price£14.00
Talk4 Podcast Mug