AI in the Sky | When Drones Think For Themselves

I really want that Jetson future you know I want to be able to jump in my flying car and go somewhere the beauty of this technology is that we can extend our human limitations yeah and and Nadia will be our Proctor if you may and this is the reason why geo fening is not working right hey hey this is air Mobility show with my co-host Nadia P thank you Amad I'm so excited to be here today so Nadia one of the things that I was really concerned about is you know we put the AI in the blue in our logo and I keep talking about Ai and know we should have more AI that what I was saying exactly we never really got into it deep right never yes all our guests on the show were talking about the Advanced Technologies in the um am industry but we never really discovered AI on its own so let's talk about it today we have yeah and we have been talking about drones and eveve toes and passenger drones and so on so I thought who better to bring on today then my own friend Mr yanush from atos we work together and he's also an AI researcher scientist Professor genius guy so welcome yanos how you you doing man thank you thanks for having me here it's uh really uh great to to to see you both uh yeah I'm the I'm the AI Enthusiast you are the aviation Enthusiast and today we'll try to make a show so let's try sounds great sound great so yanes could you just tell us more about your background in AI just for our listeners to understand what to expect well I don't like to talk about AI because uh it it became hyped overhyped so traditionally I was doing basic statistics basic mathematics and this is my core competence and of course as I'm in the technology industry I started in telecommunication industry I was working on huge networks optimizations programs and so on so the fascination of different algorith datadriven decision making and so on cames from that but as usually in our careers there were different routes and finally I was also able to implement exactly the same uh techniques uh into workplace management and now of course as we are facing the new era of generative Ai and this artificial general intelligence approach out agentic Ai and everything like that we started to treat AI as a truly partner in in in our operations on the same level as individual collaborator nice and and I wanted to you know bring you on today um you know the implementation and the benefits of AI in the aviation sector in the Advanced Air Mobility because CNS is always a big topic safety is a big topic surveillance is a big topic so you know I thought we discuss all those uh elements and uh pick your brain on what's happening in the industry yeah and and Nadia will be our Proctor if you may and she will keep us not to get to you know uh we will we will but she'll pull us back unfortunately the one to I'll be the one to be like wait wait wait can you please explain because this is a bit too uh too smart to understand uh uh for my brain and uh let's just um yes let's take it easy here uh and so on um but yes before we jump on this exciting questions uh uh maybe we just can talk a little bit about AI applications the use of it that we are seeing in um in general and how we can uh transform the Air transport with ai ai per se is not well defined term because everything can be AI these days so instead of that I I would prefer to talk about autonomous operations where decisions and operations per se are executed and supervised by nonhuman entities so let's collect this so this will be one group of systems and then of course we will be talking a lot about decision support systems different kind where Still Human is supervising all these decisions right so I I would separate these two levels of AI when we will be discussing because the first one is truly the modern one uh something that we are discussing today so the Breakthrough from I don't know uh Transformer algorith and generative AI allows us to think of implementing autonomous operations at scale while the second group that I described is present in our lives since 1950s 1960s I would say right so so of course there is a bridge there is a link between the two WS but those are two completely different WS no this is exactly the topic that need to be explained and and good to have you to explain it because the concept of neural network has been there for a long time I mean when I was doing my degree we also had neural networks and prologue and lisp and you know that kind of AI has been there yeah no that's that's very good um uh idea let's uh start in that format what's the first thing that we should be talking about when it comes to implement imp menting AI in the aviation because you know there's a man in the middle concept but then there's also the concept that AI should be completely in charge because if you are putting man in the middle and man is making decision then humans can have a proclivity to make emotional decisions non-informed decisions and delayed decisions so how would you distinguish these two how should we get into quite interesting topic you touched is this uh uh men in the men in the center approach because it's very similar to what we observe in any AI implementation in any workplace right because the the ethical discussion that we have in workplace is do we prefer gut filling managerial decisions about hiring uh uh promotion you know that kind of stuff or objective data driven algorithmic processes procedures that will be autonomously driven by AI of course life is not black and white always there will be different Shades of Gray so we are not talking about absolutes right so let's say this there is a spectrum I always say that AI is about spectrum of possible solutions so of course what we are living or the era that we are living in is that we are starting to shift left this operations from Human Centric into AI Centric or the system or the network of systems Centric right so actually because when we talk about AI we apart from algorithmic approach which we described just a second ago we have to talk also about the system approach right that this is a combination of those algorith right it's never it never works in isolation so if you look at I don't know T GPT or something it's very complex system under the hood right of course the large langage model is an engine this is a system cross connected system and the same I believe is relevant for your scenarios in ivation there will not be one system to rule them all yeah even if that would be our ambition for I don't know 50 years or 60 years I remember centralized operations centralized systems this discussion but we know that the more objects are present in the network the more autonomous the network must be because otherwise it will be completely unmanageable so the same applies in telecommunication in banking in aviation I'm pretty sure this is quite standard approach uh defining this architecture what I believe we will be living in in the future will be actually the uh the system that will be retrospectively of oversighted by people but retrospectively right so the operation will go autonomously and we will do something like system alignment right from time to time and I think that's uh the foundation of a gench model as well right I mean the idea of having number of solutions and systems and models which specialize into performing a specific task and that way they can do that task very well they become an expert systems in a way and I think that's that's really the future taking that forward I remember have a some research done on what's available in the market so let's let's discuss that conversation this is my nature I usually try to base my discuss around some kind of Frameworks some kind of scientific approach yeah so yeah let me sh let me share my screen and while you're bringing up that that thing I just wanted to add one more thing datadriven decision making one of the thing which is also lacking is that data we probably don't even have enough data there are two schools one school is saying we don't have enough data yeah and the second school which I am in is we have more than enough data we have too much noise yeah that's true from statistic perspective relevant data this is a problem right so we don't know what are the relevant data for the decisions and for the processes we want to automate yeah this is the issue not the issue about the prescy of any data any data we have too much but I agree with you the availability of relevant accurate uh data this is this is a huge problem especially that in uh Industries like Aviation human cognitive capabilities are not enough are not enough because you will be relying on in infrared U Vision you will be relying on different spectrum of sound when it comes to location or something right so we need to mix and this is a beauty of these technologies that we and the world that we live today that we can extend our human limitations human cognitive limitations with this technology right so imagine living in the world where you don't have rather right in in aviation no one can even imagine this these days right so R is a critical technology present everywhere right and how this goes this goes because it implements something that humans cannot do by themselves right so we we we are not able and we need to rely on this technology is there anyone D in dep of this of course we have planes with low Spike room not visible towards rder and so on but let's talk about Civic Aviation and in this case I believe R technology is one of the key uh safety mechanism used in the world yeah and not only that enhancing the human capability right because uh you probably saw that research paper about radio Vision where they trained an AI model to start learning from a camera and then the other input was radio signal your normal Wi-Fi 5G and stuff and then it was trained to have radio Vision so if you take away the camera it can still see everything not only that it can see everything it can see behind the walls because as far the radio signal goes it can see so you know that's that's like a fascinating adaptation that can it's fascinating but I'm what I'm trying to say is not something as revolutionary as we would think we live in this world for at least 40 years from now right at least so means that we have time to adopt of course the technology pace is faster and faster right so so we have to adapt faster and faster and we know that humans can adapt only to the certain limits so of course our ways to work where is to process information needs to adapt faster right so this is the problem that we are living in today that our traditional educational system traditional uh learning approach is not enough for this future uh needs decisioning TR traditional you know compliance and Regulatory processes oh God don't yeah I I like to talk about very simple thing which is most of organizations when they prepare process documentation they do what I call isol like documentation which means tons of papers uh thousands of diagrams workflows and so on but the problem is but this works only on paper all right let's get into the aviation use cases right um you were going to share your screen U yeah of course of course yeah let's let's get into that document preparing to this particular podcast I was thinking what Aviation business is all about because I know what AI is all about and I more or less know what Aviation is all about but I never had something like standardized um sonomy of these sectors areas that we want to discuss so I found this great paper we'll of course link it and priz the authors uh which was uh scientific paper usually this is what I use and here we have two two nice um mind maps so one mind map is traditional Aviation areas and then the outs elaborate how AI can help in those areas but also they have Aviation areas with AI applications and this is I believe something we will be uh focusing on today because it shows us what is the possible use of AI in each Subs segment of this industry starting from flight operations right how this goes then we can talk about air traffic management design uh trainings operations management Logistics customer experience and of course sustainable initiatives environment related so I think this paper gives you a very good initial overview about what we can do but what is even more important or what is even better because this is why I believe it's very good paper it's not only describing the technology landscape and Technology use case or technology approach but this human factor so means what kind of existing professions needs to adjust towards AI world right so how we we have to adjust so what kind of trainings what have Co skills we need to uh get or augment ourselves to be ready for this future right so and also this describes the way how we can improve our learning using AI tools right so this is very good paper when it comes to that because then when you will check for specific role like pilot you will understand what kind of levels of understanding on AI world this role needs right for the future right what what is even better it also evaluates current existing European landscape of different engineering schools uh and the Cula that they have how this is aligned with those future needs so I strongly recommend this paper for everyone interested about how should I align to this AI world in the future in my industry right in this my beloved industry so this is definitely the paper I wanted to share with you today this is a very good starting point for anyone who wants to get more indepth understanding of you know not just the implication of AI but also how to adopt it and how to get started on the journey we are more interested in Advanced Air Mobility rather than traditional air Mobility I think there are lots of challenges in Urban Air Mobility how drones are flying within the city how we can introduce more delivery and also more complex than anything else more passenger type of uh traffic and how AI can help with that yeah so I was looking into drone traffic management let's say so started here how different countries are approaching this maybe you have heard that recently DG DGI so the major manufacturer of drones they uh dropped the idea of Geo fencing so no Geo fencing in in the erron because they said it's a user accountability and responsibility to to to restrict this zones and that actually led me to some kind of further research and uh I was trying to understand better how we can manage this at scale yeah so if geofencing is not the way to go for different reasons I'm not going to elaborate the reasons because they can be political they can be societal and so on and so on so don't justify but simply this is how it is so is there any other way and of course one of the way is let the drones be self-aware yeah yeah so if they will be self-aware specially uh self-aware where are they they would be able to behave as expected I would say this is like human alignment right when we know what is the context we can dress formally we can dress less formally and so on and the same is with those machines right if we'll be able to give them self-awareness self spatial and contextual awareness about what is the kind of operation they are doing and where are they in the space right they will be more or less strict in specific operations right so if there is a lot of space around me no human no conflict settlements nothing like that then of course I will be less strict about the policies right because I can go faster I can go shorter I can do some shortcuts whatever right but you cannot do this centrally right you cannot do this from the system perspect you know this governmental system perspective very difficult to do that yeah the communication between the Drone and the system would provide too much latency so in this operations we require very low latency operations right so this is like uh autonomous vehicles and everywhere right car whatever right those are nanocs that you need to process the information from sensor uh through the on device brain right whatever the brain is human Edge AI capability within the device exactly from my perspective when we talk about this Modern urban uh air uh Mobility this is also the way to go decentralize the system provide on device AI H AI capabilities now what is h AI of course so what we have we have three for the lay people you know this is why Nadia is proctoring very good so very question what Ed AI is no no no very good question because what we have when you look at the moment at Solutions like CH GPT this is centralized AI so you have something in the cloud one big brain where you type in and then you get responses from on the complete opposite you can do the same on device so I can take a smaller model but put it on my PC or my laptop mobile phone whatever and this will be on device Ai and what we are talking Ed AI is a combination is a hybrid approach of the two so means we will have a local computational capabilities about the self-awareness autonomous operations but they will be supervised and could be override by the central system that comes from the cloud is this more clear for you yes that makes sense yes awesome Aji is is probably a good solution in that case and uh I think what you're describing is and immediately I'm thinking about it because uh of the research that I have been doing if you have some kind of like an ethical uh edge control Edge AI to you know learn the etiquette of the airspace you are in we don't have to worry about the Drone or the autonomy to to or you know some some level of autonomy right we don't have to worry about controlling the the device the vehicle in that specific air airspace like you said you know if it's open field there's no controls no humans there's no property to damage there's no risk you can have full control and you do whatever you want but as soon as you are in a buildup area you have property people then they already have the the ability to understand okay now I am in a congested area I need to be careful because I can damage property I can damage and even if the operator is trying to force it to do something naughty like spy on somebody it can kind of like say okay you know I I refuse to do that because I'm not comfortable doing this thing right or even more restrict when you are in a vicinity of let's say an airport or in vicinity of of where the paths are of commercial flights and it can completely block itself because it knows now my point in time and location where I am on the on the coordinate that I am I'm getting closer to the airport even though my operator is trying to push me going going I I shouldn't because this is Class A airspace Class B airspace or I'm I'm going to encroach on something right yeah that's why I'm I'm showing right now to you how drone um navigation is managed in Poland today right so we have the Fantastic polish Aviation agency that is steering all this traffic zones and within the traffic zones uh we are ma or they are managing access toward specific sub zones the problem with current model is on only that and this is the reason why geofencing is not working right this are Hardware based gensing because especially in within urban area actually we should restrict all the Drone traffic right because all these zones that you see here they require some kind conditional access right so you need to ask and get permission to go go through this Transit zone so on especially on some altitudes you know this this is more complex but I'm pretty sure your audience who is in aviation knows this better than me even but what I'm trying to say that again we are putting here our two dimensional human cognitive limited perspective towards multi-dimensional Dynamic Aerospace zone right because what other dimensions we have we have of course attitude which you don't see here right so altitude is a is a key concern and more important time but time in different dimensions so past what happened what events happened around the similar or something so we can predict some things from the past parent so what objects are present in the space at the moment and future so my prediction about future traffic right so what will go on different and then multiply this with different altitud is uh one more element in the future is no temps so you know there is you know restrictions on area in a certain time and space which is done by theps yeah yeah but what what I'm trying to say is that this will be the the calculation of rules of Entry towards specific zones will become more and more advanced advanced means they will go beyond what we are thinking about okay this is a time frame where the object will come and will leave yeah because there will be when we talk about Urban Aviation it will be like um car traffic the object will be here always so there will be no nothing like three traffics space I will not be able to to provide you can go or you can't go it must be automized if the Drone is self-aware um as you've described it it will know by default in its default settings where the fly zone and where is a no fly zone whereas right now the drones are um piloted um remotely and basically can enter any area so that is the major difference when we look at implementing the AI technology what is even even better we will be talking about swarm uh approach so means each drone will also be a it's sensors they will be a source for other other drones um operations right so means they will communicate between themselves without human supervision right so this objects will talk each other in the language that only this object understands right right so they will be sharing its own signals sharing its own decisions that they made right so this is the same what we observe right now when we try to optimize City traffic with uh autonomous cars right when we put human it's much harder to model this traffic when there is a human with its unpredicted behaviors then when it would following the rules exactly this is but this is the beauty of humans this is the beauty of humans we are unpredictable that's right but from the network perspective that's why because this is Network mapping right some kind of uh we are trying to manage Network and this network especi so imagine how many problem we have right now with introducing autonomous cars right autonomous vehicles where we have fixed roads this is a huge simplification over what we observe in the air in the air we don't have a strict roads we have channels we have of course something like a corridors that we can fly and when we'll be talking about Urban um Mobility H are we going to recreate traffic lanes and traffic lights this is what we discussed some time ago with do you have that picture I would like to show that picture I have Cy unfortunately but yes I was trying to generate that was very that was really funny you know I was trying toate the sky and then in the sky there was like a traffic light what you're talking about the the the Gap in that implementing that swarm capability or the communication capability which is also by the way very key factor of the situational awareness so that you know you know who is be around you behind you underneath you flying you know in that soup of airspace if you may around the urban environment I hypothesize a design of a mesh Network for exact that problem in my paper as well uh but that paper actually uh highlights the need of a mesh capable um cheap and fast real time neartime ability to pointto point communication and if we want to centrally if not even manage monitor it the capability for that Central management system to be able to also view that same data and I think the edge computer Edge Edge AI completely um fulfill that topic that's a really good and and the perspective that you just said right now so the conversations that we've been having uh among people who are not AI experts and in a experts have been saying that it's more difficult to train car cars to become autonomous drivers because you are restricted to two Dimension and if you open up the third dimension it would it would be easier and in contrast what you're saying it will be even more complex and I think truth might be somewhere in between because you know we won't know until we try it out right uh the yeah the one problem that we will have in this uh approach is positioning of objects and navigation of the object because with road navigation our GPS navigation is good enough right so the Precision and everything is good enough when we will be talking about multi-dimensional so this multi the multi-dimensional navigation we cannot rely on existing location Technologies right that's very true that's very true and for me this is also something that we have to invest as a society more because we still rely traditionally on what we call Elian uh geometry so GPS everything works like that and I strongly believe that we need to invest in nonan Geometry to solve this issue and you know what this you know what the solution might be and again I'm referring back to my document uh one of the things that I hypothesize is we need to incorporate in the GPS solution don't drop it or throw it away but we we include a Quantum based navigation system it's not really it exists right now it's not really a theoretical thing and the re the beauty about Quantum based navigation positioning system is the positioning system it is accurate to 2 mimet uh 2 cm accuracy in a three-dimensional space so you know but but the thing is no solution exists so someone have to pick up and incorporate into the aviation equipment like uh people Country Companies who who build that like Garmin and honey and so on I'm when we will think about it at scale again right so having thousands of objects in the space of I don't know one kilometer right so something like that right different objects small object larger objects Transit objects different altitude uh time so even if this is a three-dimensional uh spatial space adding time yeah time Trends and time predictions and other stuff will make it even more complex yes it's the fourth dimension right yeah yeah but what I'm trying to say is we have right now linear navigation right we are doing you know linear navigation point point navigation right so we will have to go towards what I call Arc based navigations right we'll be counting arcs instead of lines arcs between objects AR because only arcs will give you this additional Dimension perspective right yeah and this is what is what makes it completely different than car operations right because in car we can stay with linear algebra uh Le mathematics I told you Nadia not mathematics will have to be here let's bring the mathematic so it's fine it's fine let's teach algebra unfortunately living in AI without understanding the basic principles of MAF is not possible but this this is what what what made me think how we can improve the current state of uh technology and use this AI use this Aviation AI in aviation use cases to improve prove also a general science right because this is this is what we are missing for we we focused on this um existing linear algebra approach and we Excel here while thinking about this multi-dimensional navigation systems I was discussing this with some scientists and we have something that we can say scientif science fiction approach right because when you will look how we fought our um yeah you know Star Trek ships are flying or something this is exactly what we want to model today the same approach we are not talking about something different of course they were talking about uh traveling with faster than than light speeds which is physically not yet possible hopefully with Quantum Computing and quantum physics go go yes yeah maybe in sometime from now or or we can bend time and space that would be good as well right let's think maybe you know sorry but this is also something that we need AI for because our our our imagination is limited yeah that's that's that's true and I think I think that's where a big huge problem lies right because masses are simple and they are confined to what theoretically they can think of and this is where also the aviation problem lies because the aviators have been aviating for a long time in the way shape and form that's only they what they can think of you know like uh we were talking about the concept of uh threedimensional four-dimensional just you were explaining I was talking to somebody in aviation who you know knows how planes fly and what routes they follow and how they follow it and you know how they make that work but you know that's not transportable into the advanced immobility model if we are going to go advanc we can't just modify what we already exist and stay in the box we have to think outside the box and start thinking about we cannot have the reporting points the way we cannot have the Airways we can we have to have dynamicism and that Dynamic approach to be able to uh really use the three dimensional in fact four dimensional like you said the fluid that we are it's not only the the location in three dimension but the time and point in time and space as well right so it's it's it's way more complex but that would be the breaker of this problem we were talking about drones and the ideal picture of drones and them being um self-aware situational aware um but from the security standpoint uh there are so many counter drone Technologies being developed veled uh to actually go against those uh drones that are potentially so how would this technology that we you just discussed help us as general public to be less scared of potential drone attack very good question and I'm looking at this yeah so of course we can start with military comparison because we see the war in Ukraine this brings a lot of advanc ments with all the cruelty of this war and all this you know suffer that the war brings but unfortunately this is also the old Tru that war brings Innovation and especially when we talk about drone rones and counter drones and everything around drones this is a huge when you will look how much advancement was done over last two years right three years something like that right in this area it's uh it's huge uh when I've seen the latest trends so the Russians are going back to fly by wire right so they attach the drones with I don't know 60 kilomet long fiber optical fiber yeah and they fly like this right so if you will ask me how this can prevent the only thing will be how we can identify objects in space so what will be the trust Network to authent iate objects in space right so which object is uh authorized which is not you know this what we get from crypto business all these Technologies for blockchain Technologies everything like that this will be the way to authorize objects in the space right in the future so definitely that kind of approach because then this is the how you define friend of four and counter measures means all those rifles that you can disrupt elect elonics on board Electronics they will be your last defense they will be the last to be used right so no one will go with its own rifle to shoot down the drones right so instead of that what we want to have we want to have a strict ax Network or the space access protocols right and authorization here and this is what I believe we can use this blockchain technology for to have non-centralized traffic management approach right non-fungible non non immutable encrypted in a way that cannot be imitated except for whenever Quantum Computing breaks it then you know yeah this is I believe something we can we can we can think of and then there will be so like today you have transponders on any objects right so any flying objects they should have something like transponder right so the same here right if if they will not have authorized access they will not be allowed and then of course you can Target them at the Spectrum of counter measures right so so not to to I don't know make unavailable the full space but very targeted uh precision based um counter measures friender fo those bigger passenger drones that uh the air taxis the EV tools that are being developed right now do they need to keep into account that they there will be other flying object that will uh identify them at a friend or a foe and and there are bigger scale of things as Government security and things like this so from the ideal standpoint it means that absolutely every drone needs to have this technology inside of um of its frame to be identified as a friend or a foe right be part of the ecosystem um in the urban airospace as we discussed this modern traffic manag mement will not be light based like we have right now but right because right now we have more or less right again the same red green and yellow lights everywhere present even our air traffic control right so we are giving access or not right so simply like that so you have approval or not in the future you will not behave you will not rely on this uh rule based traditional uh conditional based Access Control instead it will be as I said you will have have a authorization or access and authorization controller on board and then it will be encrypted with all this modern I don't know what kind of encryption will be invented with Quantum but so far the one that we believe is the most uh safe is the blockchain approach right so what we can do here is is the most uh trustable because there is no one source that you can compromise and you can go into and you can say ah now this is allowed right so with all the centralized system this is the biggest problem that let's say I will corrupt the administrator the human is always the the the most vulnerable you know it's like the movies yeah I'm going hack into it and override it and change the P you can't do that anymore because it's not centralized so this is what we want to or what we envision in the future of this uh access yeah and I think who job it is sorry just I'm just so curious so whose job is it to to make this happen well of course there will be a lot of no no no no no uh you know U again my Rebel my Rebel uh soul is not letting me say that this should be government that's the Polish way man go for it no it's a Well there will be communities right so we will we we have communities and these communities will include government bodies regulatory bodies uh Airlines uh manufacturers but also operators of those um let's say autonomous operations right so we need to decide how we manage space in the future right are we going to have the same uh governmental perspective towards air Space Management or this will be commercialized right so like you know Highway management right some countries have them private right so commercial ones right so commercial model and I can imagine about the same rules right and principles that you will have a specific areas that will be uh restricted by governmental bodies like military bases airports everything like that while some other areas will be managed uh so what we how we started right with this less strict uh Access Control where this can be even I can think about making them commercially available for this Aerospace management perspective right so imagine you will have an operator that will make sure that all this Access Control uh mechanisms are validated in his space right according to of course policies and rules so you know you have the adsp system like you said the transponder and you have the radar system right now and what you do is you have a transponder transponder have capabilities and then there are service providers who are responsible for providing that data that is uh used by all the commercial airlines and airports and anps and you know so on and so forth and what it does is each transponder identifies the aircraft and then also transmit data like you know altitude you know this and that and speed and blah blah blah so now just imagine that and imagine Bitcoin and Bitcoin basically is a cont contract which is computed and uh then it's issued and that database that contract database doesn't actually exist in one location it is loc it is you know not decentralized there's no centralized thing and it's not controlled by anyone then so basically once you have a token and once you have something issued and that can be the basis of the encryption or identification or whatever how way you use it it will be a combination something in the middle of what adsp does and how we use the crypto um technology that is creating these you know cryptocur currencies today and implement it in a combination of that kind of thing I'm very happy how we moved from traditional AI uh towards all the spectrum of technologies that we need to make this uh dream happen right so and that's what a is all about Advanced Air Mobility is inclusion of all of that stuff you know uh what we'll have to do is come up with technology that uh help securing your flow of data help securing that identification and authorization process and then also verify and quantify the information transmitted because what we also don't want to do is a drone which is on the network and start sending wrong information and start to cheat you I'm not behind you I'm 10 kilometers far away but then it's spying on you by being right behind your tail yeah so you know there's a lot of things to think about and all of these are the uh technology aspects that am have to incorporate into the research and the workflow and honestly this is why I'm starting the open am organization the as an open source something that everyone can collaborate in and uh do and yanish you're definitely invited man I mean you have to you have you have to you know help me out with that because that is the most complex thing that we are taking on to make this happen like you said you know to make this happen there's a lot of new things that we'll have to look into and the purpose of the organization is to research and contribute and talk about and you know voice these concerns I I I really want that Jetson future you know I want to be able to get up and you know jump in my flying car and go somewhere or at least call a flying taxi come pick me up maybe not from my doorstep but at least you know down the wrong road at the exactly you know my neighborhood workport come on you know I would like to see that the future excites me I know the for practicality and all that but this Urban Mobility is even less about passenger traffic but basic Supply traffic right so we will have a lot of drones different drones right yes running with I don't know medicines supplies uh this grocery stores imagine whatever you can right so this things right and this will be majority of the traffic human Transportation yeah it's a fraction it's a small fraction of this right compare this to the how Railway works right so you know rail works right so you know when you go by rail so the the the you know the passenger traffic is I don't know 20% maybe of the overall uh traffic but the commodity flows resource flows information flow this this will be something raw materials and all that sort of stuff yeah no no you're you're absolutely right and and I think there are two two things in this one is that um you know I think the use cases of and and I I like to talk about this in every single podcast and I know my viewers might get you know stop talking about it but no I I I I I sorry I can't because I think it's such a important use case you know because it's quality of life Improvement things like Emergency Services air ambulances you know medical deliveries organ transplant deliveries blood transfusion deliveries this is so important and a use case that can be implemented immediately straight away because not only that there is a need and and we can but also the technology is there right now and we have discussed in other podcasts as well I think the human transportation technology is not quite there yet so you know the Battery Technology is not there to carry heavy loads further in the distance it will be smaller so you know for now there's a technological impact that we can't but also you're right you know the bigger use case is deliveries you know and I imagine if I'm living in Dubai because I do have an apartment in in Marina downtown uh sorry not downtown Marina in Dubai on the 17th floor I would like to be able to order my pizza and get it delivered on the balcony I mean that would be Fant that would be fantastic right fresher it will be crispier it will be hotter when I'm talking about when I'm talking about deliveries I'm thinking about package or this uh of package driven priority traffic management yeah right because your pizza can wait over somebody's blue right so let's agree at least on that right maybe completely agree completely yeah so so maybe it will be less uh crispy but someone uh will will will be healthy right so I'm I'm happy with that absolutely happy and with this case but but why I'm saying so because right now we have privileged vles yeah yeah so fire ambulances police cars right in the future with Urban Mobility yeah all vehicles will be equal only the load will determine the priority and this was uh I think the ep the topic for another episode to be discussed we can go on about that Logistics right so because for me this is completely related with Logistics right how we will manage Logistics how we'll order those uh uh objects let's say this uh UNM objects what they will be transporting and so on right so I think it's a very good I think that's a that's a very good point as well right because uh we have like right now Amazon drones and uh we can say okay you know this is an Amazon drone flying so it's probably not important but what we should be looking at is what is the payload right exactly if uh there is a drone which is XY Z identified by a number doesn't matter what if it's Amazon or whatever if the payload is let's say category emergency they have priority over everything else traffic right and I think that then translate into um either it's blood or organ but then also like it's a emergency response for for uh bringing a doctor to the emergency because we cannot create right now we are limited by the technology battery and propulsion to bring an ambulance so what we do is bring a Hos a doctor to the emergency where the First Responders are there and probably by having a doctor there we can save somebody's life you know uh rather than trying to transport that person to the hospital hospital to the doctor one use case for example and you know another thing like you know a drone with fire depressant um technology you know that can get to the fire while the fire brigade is you know is on the way I'm not saying just completely remove it but then there could be some drones that can go in and spray some fire suppressants very very quickly and the impact of that fire can be reduced even if you can't completely uh you know uh control it with the drones but at least you know so I think that's the first response right the first responding is also have a huge impact and I think it's a very very important use case yeah for me it's not about AI it's all about those use cases we discuss right and then AI is something like a brain that you put inside right so of this right so but use case determines the topic for us I think that's that's a very good point right AI is an enabler just like every technology is every technology is is an enabler to perform a function that we want to improve or make it better and make it more efficient fantastic thank you for your time and uh very very insightful and amazing structured conversation to help people understand what the AI is what the role of AI is and how things can improve roov and also to the audience who tune in every day I'd like to thank you for your um coming on and watching our episodes we have had amazing fantastic um growth very very quickly and I can only thank you from the bottom of my heart that you tune in and watch the episode and support us uh so please like And subscribe and we'll see you in the next episode see you bye-bye

2025-03-25 22:18

Other news