# CEO to CEO Fireside Chat

- Event: NestGen ’22
- Published: March 3, 2022
- Length: 44:53
- Watch: https://nestgen.org/library/watch/CghijFQsMaU/

> The transcript of a recorded NestGen session, for reading or for use with an AI tool. It comes from the video's
> captions; automatic captions can mishear names and terms.

## About this session

The world came to a standstill during the pandemic. On the flip side, it kicked off a myriad possibilities for drone autonomy. Find out how.

## Transcript

**[0:13]** hi everyone this is jason sansouzi chief architect of juniper unmanned it is my pleasure to be moderating this session the ceo to ceo fireside chat here at NestGen 2022 here on tuesday I hope you've been having a great session so far a lot of networking it is my pleasure to introduce this panel on the effects of coven 19 on drone autonomy we're going to be diving deep with these illustrious ceos and without further ado let me just have them introduce themselves briefly and then we'll dive right in to the questions so let's go and go around the horn let's start with gemma hi guys I'm chamber orcok I'm founder and ceo of skybound rescuer and we specialize in the use of drones for public safety um I'm also the director of air operations for lowland rescue which is a search and rescue charity here in the uk with 34 star teams and we search for missing vulnerable people on behalf of our police forces and

**[1:18]** I manage how they train and use drones and interact with the tsar helicopter services excellent let's move to dawn zoldai thank you so much hi everybody dawn zoldai here ceo of p3 tech consulting a little bit of background let's see 25 years active duty united states air force as a licensed attorney and they call a jag and two years general counsel for the air force and then one year on faculty at the united states air force academy so my business I do have many many things jason is well aware of this because like him I have my own podcast called the dawn of drones every wednesday night mountain on drone life tv on youtube and uh write a lot and I teach at two universities so uh that's kind of the big big overview super busy no this is a great panel and let's move last but certainly not least our host for this particular event the team gupta

**[2:20]** thank you jason uh so hello everyone I am nitin gupta I am the founder and ceo of FlytBase so as many of you know uh FlytBase is in the business of building software tools software platforms uh helping our customers and our partners work or not build on drone autonomy so our focus has been drawn upon me how to enable autonomous drones for various commercial applications and we partner with know various other various other companies in the drone ecosystem to provide the right solutions so like for example we have partners with we are partners with drone manufacturing companies the docking station manufacturers as well as drone solution providers and drone users and finally together we are able to provide the complete solutions for drone economy so thanks thanks jason absolutely all right so a couple house p keeping items we're here on the main stage if you have questions you can go to the q a section and ask in the backstage staff will see up those questions for us but we do have our own list of questions we're

**[3:23]** very excited about them so without further ado I'm just going to jump right in but first actually one one thing real quick if you want to learn more about these panelists go to their linkedins connect follow go to their podcast subscribe smash the like button all those things this is very important we want to get our messages out there we're all here to elevate the entire industry so let's dive into these questions okay so this is a question for everyone and it's a bit of a softball so we gotta start easy start easy then we're gonna get way deeper so what are the applications sort of the low-hanging fruits that are the most appropriate for full drone autonomy and let's let's uh we'll go backwards around the horn let's uh let's talk to nitin first sure so no again when it comes to drone autonomy of course it becomes very important to ensure safety and reliability of the systems uh safety of course is the is the kind of the most important thing that we that we care about for autonomous systems

**[4:24]** because in an ideal situation we are not expecting any human intervention whatsoever so the systems have to continue to operate in a way that they are not causing any harm or any danger to anyone's life or property now of course again in order to ensure the safety or in order to increasingly build more and more sophisticated systems that are capable of ensuring that safety we have to first start with applications that are inherently not very dangerous in nature like for example using drones for say mining sites and these mining sites are quite away from the crowded areas or from from the population and that is why it is easier to get permissions or convince the regulators to allow us to do autonomous operations for such applications versus using drones for say delivery in an urban environment which is inherently crowded and it's a very complex airspace you have other manned aviation in that airspace as well so there are too many considerations that all know have to be taken care of at the same time and that

**[5:26]** is why it becomes very complex so what we have seen us is that most of our customers they are first starting off with applications that are away from the crowded places and where it is easier to you know get those sort of permissions or convince the regulators that the entire operations will be safe even in case of some mall function uh it will not be catastrophic it can still be managed and then as people start using or deploying such applications they are able to build their safety cases or ensure that they they have proper sops and proper systems in place and they are able to generate enough data from from that use case to convince the regulators to increasingly get permissions for more and more complex scenarios so so that has been our uh kind of experience so far uh that makes sense and thanks to tim let's uh before we get too far into the regulations see that I want for dawn here in just a moment don what do you think are sort of the lower hanging fruits for full drone autonomy so we're going with the 4ds here jason right the dull dirty dangerous and difficult

**[6:33]** I mean this is exactly why we love drones we love robots this is why they will change society so when you think about some of those uh you know right off the top I'm thinking about autonomous bridge inspections like we're already seeing in the united states in north carolina statewide due to right uh skydio's um statewide bv lost waiver for that purpose I mean that's that's a perfect application um you know I I also think of the choctaw nation of oklahoma and some incredible work they're doing out there I remember one of the very first articles I wrote for inside on man systems magazine was about them and they were literally dropping these giant boar buster feral hog traps from a drone and you know I mean when you think of dull dirty dangerous difficult I mean it sounds silly but I mean if you have to do that by hand and those animals are very dangerous people get injured all the time

**[7:36]** you know trying to do things like this so they're dropping the traps they're also dropping I bel I believe feed into them as well so you didn't have to go out and feed them so um just just two examples I mean there's hundreds more any job you can possibly think of that meets the 4d criteria perfect for autonomous drones great now I completely agree don absolutely the 4ds are very well suited for full drone autonomy and we'll dive into a little more granularity on that in a few minutes gem over to you I'm sure you have a different perspective yeah I think for autonomy um repeatability is important of the mission and unfortunately public safety we don't get much repeatability no two emergencies are the same I mean if we are getting the same emergency many times over then um perhaps our prevention needs a bit of work um so something like searching the challenge researching is there's no two searches that the same so it takes a long time to gather data understand the key variables and their dependencies

**[8:41]** work out what works what doesn't in order to get to automation let alone autonomy um but there are other applications that are slightly more repeatable in my world um so for example um an application I've been looking at with the coast guard is uh patrols aerial patrols of high risk areas so uh multiple times a day the drone will patrol a high-risk patch of water to try and find persons in distress or at risk um the mission plans are identical the area we're covering changes slightly because you know rips move position they unfortunately don't stay in the same position all day um but it's very repeatable however there are still variables that impact our performance that will be different every time things like weather um you know the the classic um weather can impact visibility amongst other things so one patrol we might be able to easily find that person and then the next patrol uh only a few hours down the line perhaps it's high

**[9:42]** c state and no longer do we have a high probability of detecting that person so in those cases what we're doing is um we are scoring those variable factors those factors we can't control so that we can say that patrol uh was done to 50 quality and then the second patrol was 70 quality so that you know we don't have to have a person in the loop assessing the performance but we still have an indication of performance because that is very vital for public safety a very unique perspective yeah I would say we don't want any emergency be exactly the same over and again so that makes sense um so don we we talked a little bit about this one's questions directly at you uh how has coven 19 impacted drone policy and regulations well you know jason I I think oftentimes necessity is the mother of invention and what we saw during the covid19 pandemic which continues to persist is this idea that we you know social distancing is is something that is desired and the best social

**[10:47]** distancer is a drone uh and so what you saw I think really from march of 2020 through now was this explosion of use cases in particular uh you know drone deliveries of things like covet testing kits food rule areas and I get to this I'm telling you about the use cases only because it's relevant to the regulation and the policies that we saw evolving in so far as in the united states you know part 107 got to stay within you know visual line of sight we saw a number of beyond visual line of sight waivers getting approved for this very reason and in fact jason just this morning I was surfing around the FAA's website which by the way is spectacular um if you go on their part 107 waivers issued link you'll see and you put in part 107.31 which is visual line of sight you'll pull up

**[11:49]** 85 entries okay for for beyond visual line of sight waivers essentially uh and the most recent one was just 2 16 20 22. so what was that last last week last week to iris automation um and you'll see the the companies that are getting these what they're using them for um so again what we see here is some additional kind of consideration uh you know and flexibility on the part of the regulator at least in the united states the other thing I'll say about that of course is the beyond visual line of sight aviation rule making committee uh the you know the bv loss arc kicked off last year as well in earnest and we're just waiting to get some word on that and so and then the final thing I'll say is kind of getting away I guess it's still related to beyond visual on site but also connected technologies I know that's natin's like

**[12:49]** area here but um you know our beyond program the faas beyond program which is that continuation if you will of uh the integration pilot program the the research and development aspect of what they're doing out there is not just looking at beyond visual line of sight but the the critical technologies like detect and avoid um you know that the FAA is going to need to certify right that these are legitimate these work and so all of this ties together and that's what I think you're seeing on not only the regulatory and policy front but it just it all coalesces together including the r d front and then what we saw as a result of covet right absolutely like I said yeah I couldn't agree with you more mother necessity of invention we also know that there are other worldwide people are looking at this differently so this was the u.s perspective worldwide there's definitely a

**[13:50]** we're all trying to trending the same way in bringing the industry up uh so thank you don appreciate that so next question is four ninten so drone autonomy it seems super cool right there's a lot of things we can think about when you're thinking about your autonomy but let's talk about the realities right what are the realities of drone autonomy let's really dive into this and how are the problems with full drone autonomy being mitigated today right so jason as a as a drone autonomy company what happens is we get very excited about building cool technology right so we are always excited about ai machine learning and making swamps and no intelligence systems and not so many things right and it is of course you know very interesting and exciting to be working on those things but what happened is uh over last couple of years because of covet um we really we realized that there are so many simple easy problems for which we still do not have a a good reliable solution which is affordable which can be easily deployed

**[14:54]** and used by the masses and that is what really struck us and that is why you know we decided to focus on first solving this problem of drone docking stations or helping uh know our customers uh to to be able to conduct these missions without having a person on the side or being able to conduct these mission setting remotely right so of course no regulations are one thing but even from the technology side even from the product availability side uh it was such a big problem you know we had customers who were operating in areas where they had very good relations with the government they had the permission to operate a drone you know few hundred kilometers or few hundred miles away but they just did not have the right products uh accessible to them that they can immediately go procure and start using right so these docking station uh for example these used to be extremely expensive and this is very complex there are very few manufacturers who are building these systems and just because they were selling so few of them so it it's kind of no I mean chicken like sort of a

**[15:59]** problem that they were building so few of them so they were very expensive and because they were very expensive and people were not able to buy and help them build these at scale right so so this was the realization that okay uh drone autonomy is of course the future and we all agree to that but where do we start from and let's start by first of all enabling the customers who have the permissions who have the requirements to deploy remote systems to be able to do that in a way that is affordable that is you know reliable and it like kind of know works all the time right and that has been kind of our focus for last couple of years now so we we basically went out and we looked for okay what is available in the market and how can we help these customers who want to deploy such systems and there was like literally nothing available and there were a few uh companies that we came across who were quite interested in building these docking stations but again uh you know it it's initially it seems like a very trivial thing it's just like a little box to

**[17:02]** house a drone but when you start uh thinking of a production quality system or system that can actually be used but I know in large quantities by these customers day in day out uh it has so many complexities within and know if you start solving all those problems and building something which is production quality which can be mass produced and which is affordable uh it becomes quite a massive uh kind of endeavor in itself and that like that is what we have been doing for last couple of years now and now we are really proud and happy that now there are no more than 10 docking station partners that we have been working with and the cost of this docking stations has now come down from the range of around 200 000 all the way to just about fifteen to twenty thousand dollars so which is amazing and no it's a fraction a fraction of the previous cost or order of a magnitude less than I know what these systems used to cost previously because of which these systems are now being uh no they have become very affordable and that is why the number of systems that we have

**[18:06]** been able to sell or make available to our customers is far higher than what the no other manufacturers were able to do in last five years so so that is a big change that we see now and that is the no kind of immediate problem that we wish to solve that how do we make these systems super affordable and super reliable and make them accessible not to all these customers that's those are all great points so scalability and affordability accessibility those guide the changes that are happening on the hardware software side don on the standards and certification side what's guiding that what needs to change for standards and certifications for to get drawn autonomy to happen well I don't I don't mean to sound like a broken record here but really going back to the beyond visual line of sight rule I mean we need one right so right now we're

**[19:08]** operating by exception if you will you know I mentioned there's 85 you know waivers approved that's great but I think that was in a sheet of 3700 plus waivers so do that math and you can see the small percentage that even though we're seeing kind of a little bit of an uptick in that we're just not there yet and so um you know the research in the beyond program and um you know what people are doing real world with these waivers will inform that safety case and will in fact inform whatever that final rule is and to your point earlier jason absolutely globally we're seeing the same thing um you know so when you think about europe and you think about their use based regulation which essentially is their you know unmanned aircraft system or on crude aircraft system traffic management regulation I mean they've been exercising

**[20:10]** real world like exercising uh these technologies and their regulation uh for about two years now they just had the trials the gulf uh gulf of finland 2.0 this past fall and that's going to continue on and so you know um you know what needs to change I think um you know some regulations still need to be written the ones that are written need to be basically proven and all of that is underway um you know the other thing I'll say about canada you know we think europe you think canada you think australia um you know canada australia you know they're they're doing extended visual line of sight or eve you know vlogs and some of the rules and I that's in australia and I believe canada this this past year uh passed a um a low risk beyond visual line of sight regulation or at least they were tracking uh to do

**[21:11]** that um you know that's going to require some training uh additional training for those pilots because you know this jason you're a pilot you know and gem of course operating beyond visual line of sight is is a different deal than having something within you know within your vision and so uh that training is going to be critical too it'll be interesting and see if that winds up in the final rule uh you know or at least the notice of public rulemaking that we're hoping might come out this year in the united states so globally you know what needs to change I don't know that anything necessarily needs to change I think just things need to continue to progress things need to progress yeah for sure so jim this next question is for you um we're talking about strong autonomy from your perspective how is like the appetite for it changing I mean there's a scale right from sort of level zero to level six right so as we move along that scale how is that appetite for drone autonomy changing from your perspective

**[22:14]** good question um I think uh the appetite for autonomy um in public safety sector anyway is comparatively lower than other verticals as I mentioned before you know we don't have much repeatability um and I would say that they want it and they expect it in the future but they're presently quite skeptical skeptical about its maturity and accuracy um it tends to be quite a risk adverse sector because we're not just risking time and money you know if we make a mistake risking lives as well the lives of the public and our team members as well um so with autonomy more so in our industry than other verticals within the drone industry the system needs to be able to make decisions uh to the same standard or better than the skilled people of today and that comes back to the problem I mentioned earlier of we need that uh data and with some emergencies like a search because there's so many variables and they're

**[23:15]** always changing it's going to take time to get enough data so that we can actually learn how do we automate this and work towards autonomy um so yeah I would say that the appetite in my sector um is that it's improving um they want it uh but it's and they understand the benefits of it right response times are everything um I think speeding up response kind of is a proxy metric for um saving lives the faster we can get the faster we can find them the faster we can rescue them and the better the survival rates so we understand the benefits we want it it's just it's not gonna um we might even be the last vertical to um enjoy autonomy because of the complexity of getting it right and the importance of getting it right okay jason I'd love to jump on to that actually yeah go for it when you're talking about appetite you're talking about public perception and you know that's something we talk about all the time you know within this industry of how do we pierce that

**[24:19]** bubble how do we get the public to accept and you know one thing I'll say about the FAA when they did their drone safety awareness week this past year you know one of the themes they had a theme every day of the week and one of the themes was public safety public acceptance and when I first saw that I scratched my head I'm like what what is this even about like what is the connection but I mean I think gemma just really hit the nail on the head when she when she know you know said look the public is watching public safety like when they see a drone that public safety is using that is forming an impression on them and that's frankly where they're seeing it they're not seeing it out on a construction site usually right uh you know maybe they'll see a delivery in their neighborhood maybe but where they're we're very visible is in the public safety sector and where I think autonomy would be perfect is what we're seeing already like out in chula vista you know the drone is first responder getting to the scene quickly um you know but you think of other applications in public safety arena we don't want

**[25:21]** them to be autonomous like we don't necessarily certainly for a lethal force as an example if you've got a fleeing felon you know you can think of the scenarios that public safety or in particular law enforcement would have to you know might want to use a drone for some of those there should always be a human in the loop like there should not be you know and I think the military at least the united states you know from an ai policy standpoint has come down that way too of having this human involved you know we're not going to have drones and robots making decisions taking people out so you know it just you know the lawyer answer is it depends on what your exact use case is whether in the public safety arena or elsewhere whether or not it's a good fit fit and you want it to be a good fit because if it's not now you've just detracted from this idea of public acceptance absolutely I'm just going to add on this because I think this is there's been many nails hit on the head not just in this session but but grant earlier today and some

**[26:25]** others to talk about public perception I like this comment I would love to see some facts behind this but I'm willing to bet it's true gemma there's been thousands of worldwide drone rescues thousands I would say if you start to weigh those compared to the the off uses of oh no a drone's spying on me and it's actually spying not just they're inspecting utility poles and you know karen thinks your spine um I think you're gonna see that that that benefit is starting to shift especially as as more drawn autonomy comes up as the drone industry starts to bloom you're going to start to see more on the public safety side that is more of the life-saving more of the you know what look these are the I mean if you're really saving drives live with drones which we have absolutely proven and it's accelerating rapidly those use cases are gonna start to like dawn said they're going to make public perceptions swell in the positive away from some of the potential use cases that you

**[27:26]** know what as standards and training happen for more drone drone for more drone pilots and start to be put into these autonomous drone-in-a-box solutions some of those things are going to go away there's always going to be rogue actors but for the most part some of those things are going to go away with just good training and I mean yeah so that's my opinion I think it could be really interesting I like this topic specifically but there's a lot we could do and we have more questions tim coming to you next so you actually mentioned this a little bit earlier there's there's more than a dozen drone box solutions that are available on the market so how do we make and you mentioned making it more accessible and approachable how do we make it more accessible and approachable using lower cost commercial off-the-shelf solutions yes jason so basically that was the whole uh no idea when we started two years ago to solve this problem that what we realized is that many of the companies that started initially building drone in a box they started in 2014 2015 time frame when

**[28:29]** reliable of the shelf drones were just not available so at least if you if you if you remember like know how the situation was six years ago seven years ago it was uh no really very diff like there were just no off the shelf drones you you would have to assemble a drone to create a drone right so you would buy motors and esc's and batteries and everything and put a drone together and these drones would just kind of keep falling apart or know something all the other will keep getting broken um and that is the that was the time when these companies started and they started building the drones the docking stations the software everything all by themselves and that is what led to of course no huge investments in r d uh and rightly so however what has happened in during this time is uh with the emergence of DJI parrot autel and some of these other manufacturers who were able to really solve the problem of getting a reliable drone in the hands of the masses right so that really changed everything uh you know in the drone world as we know uh

**[29:34]** and what happened is because they were able to produce uh drones at such a large scale they were able to bring the cost down and not only the cost down but these systems started becoming more and more reliable and now because no for example mavic is sold in maybe like millions of quantities and it is an extremely reliable drone which is available for like 1000 1500 1500 or so so that that's uh that's really a game changer if you look at it right so uh five years ago six years ago trying to get a drone at the level of reliability of a mavic would cost you maybe fifty thousand dollars or something right and now because it is being mass produced they are able to lower the cost and build it uh you know at production level you know in such large volumes so the whole idea when we started two years ago was okay let's let's rethink do we really need to create a new drone in order to build a drone in a box system or just leverage what is already available in the market and work with some partners to create the right docking station for that drone and write no

**[30:38]** subsystems etc cooling system weather integration and this and that so so that that has been the approach and this approach helped us lower down the cost significantly as we discussed and it's not like half the cost one third the cost is one tenth the cost so it's a it's an order of magnitude it's a huge difference and uh that immediately opens up a very new market earlier two years ago if you look at you know there were like maybe two or three docking station uh providers and they were selling only to the government customers or no customers who could afford those quarter million uh dollar systems and so on and now we are beginning to see an emergence of an entirely new market where people are planning to deploy these drones for you know their factory sites or their mining sites and where they they could not have been able to afford a quarter million system but no twenty thousand dollars yes uh no it makes sense let me let me try it out let me start using it and if it works at one of my sites I will quickly expand

**[31:41]** it to 20 of other sites right so so things become you know very interesting uh once you are able to bring things down to that sort of a price point and make it affordable and available uh no so easily uh to a market that was underserved or not not served at all uh just a couple of years ago so so that is where we are and you know working with partners and leveraging each other's expertise uh to bring the cost down because everybody is now doing things at scale you know the hardware manufacturers are building at scale the docking station manufacturers are again not trying to sell hundreds and thousands of these docking stations instead of trying to sell only five or ten of these uh no in a month or in a year as many of the other docking station manufacturers used to do two years ago so so now everybody is doing at scale because of which everybody is able to enjoy those no economies of scale and bring the cost down and ex help grow the market absolutely now that makes sense um using

**[32:42]** mass production and having them bring to the new use cases right this is this isn't a mark niche market anymore this is now we can throw these on trucks go out to a site we can put these in ag fields we can put them in construction sites they can go all over the place they can sit on a beach and look for sharks I mean like there's all sorts of things that fully autonomous systems can do without having a person in the loop and I think this this this particular summit has definitely shown that now I'm gonna I'm gonna pivot I'm gonna ask a different question so and this is for everyone and we're going to start with gemma how far is too far with drone autonomy I like this question um and I think my answer is quite controversial given the space that I work in uh I think most technologists have a vision of someday a team of robots does everything and let's say someone's drowning and maybe use a drone to search and find that person then drop an inflatable and then um an autonomous

**[33:42]** surface vehicle comes and rescues them to shore right um and uh kind of whilst that vision sounds awesome um I think I I always try to start with the human um and how it might feel as the casualty in that scenario that is drowning then they've got something dropping on them and they've got this robot driving towards them um would that add to the distress um and also I'm a bit concerned about the reality of the situation the reality of that scenario as well um I have had to paddle out many drowning casualties I was a beach lifeguard for seven almost eight years and on quite a few of those occasions all they needed to do was grab hold of the handles on the rescue board and they would have been safe instead they chose to launch themselves at me pull me off my rescue board um I have to drag them a bit deeper for them to let go and then we try the rescue again um that is the recognized way of getting someone off you by the way I'm not just being mean um and uh I mean my point is that a drowning person is not

**[34:48]** thinking rationally um they're not thinking logically and it is often the case that a person drowns a few meters from help um in the uk uh cold water shock is a big killer um so you go into the water and you go straight into the shock state and uh you can't control your breathing um you're submerged and you're essentially panting underwater so you take on water really quickly you have rapid extremity cooling so you lose all dexterity so basically you lose your ability to swim even if you're a strong swimmer in fact um university of portsmouth over here in the uk uh tested this on olympic swimmers and even they lost their ability to swim when under the effects of cold water shock so firstly I do see some issues with dropping an inflatable on a drowning person um for them to grab hold of um I know it has worked on some occasions but it has to be able to be as reliable as a human rescuer um in order to be a viable alternative it

**[35:53]** can't just work in a few situations um and then throw into the mix um kind of these rare conditions like an example is the autonomic conflict death so it's rare but it happens it's when again you're in cold water clearly I know cold water very well too well um you go into brachycardic state which is a very low slow heart rate and then you see help is coming say you see the autonomous surface vehicle that creates a spike in adrenaline and then you go tachycardic a high um heart rate and it's that sudden conflict that causes cardiac arrest it's rare but it can happen can that autonomous surface vehicle save someone in that unpredictable event uh as well as a human rescuer can you know these are these kind of unpredictable events can robots do it as well as humans um but all that to one side if everything went to plan would a person rather be rescued by a human or by a

**[36:58]** robot you know it's a traumatic event to be uh drowning um so the experience of the casualty I think should be taken into account um so I guess my question back to uh how far is too far with drone autonomy is around how does it make the casualty feel um and can robots deal with these unpredictable and every eventuality of an emergency as effectively as a human rescuer can um my view is that robots should be supporting those human rescuers rather than replacing them um so any autonomy that replaces the actual rescue happening like by a human I think personally is too far um unless I'm proven uh wrong on those two questions I mentioned um but yeah that's my view on it that would be a fascinating study honestly like what if it looked like wally what if it was wally coming at you rather than like it's shark looking torpedo you know obviously one would terrify you and the other one be like oh it's wally

**[38:03]** um let's go to dawn I would love to get your perspective on this question how far is too far you know I'm going to say one word here and terminator no so look you know with my military background when you think of autonomous systems my brain goes right there because you know I I can't think of too many commercial use cases where it would really offend the senses when I and I when I say commercial I mean industrial right like agricultural you know when you think of all the normal verticals I'm not talking public safety okay um but when you think about public safety in the homeland domestically or you think about the military on the battlefield um that's where I think we need to draw that line and you know there's a brilliant uh brilliant woman at duke university dr missy cummings she's a former uh she's a retired um uh navy pilot and you know double engineer I mean

**[39:06]** her credentials are impeccable and insane but she's written extensively about this concept of autonomous weapons systems uh you know particularly for the military and like right now of course you know to gemma's point you know there are always going to be dynamic situations uh you know you've got a drowning person they're going to panic they're going to well on the battlefield there's all or you know in an urban environment domestically there's always going to be dynamic situations um where you know okay there's your target but now all of a sudden a little kid's running into the right into the line of fire and you know it it is that that that is why I think it is always so critical to have a human in or on the loop shameless plug jason I'm just gonna do it nitin I beg your indulgence um you know uh gonna have uh an ai policy panel at an event I'm running at auvsi exponential in orlando

**[40:06]** florida on monday april 25th um called law tech connect workshop I I mentioned this because world renowned people are part of the faculty for that um on the military side pat houston brigadier general united states army retired was one of the is one of the preeminent minds in our artificial intelligence autonomous policy uh on the battlefield and he is one of our panelists that's just one example of so many so um we grapple with these we're going to be grappling with these topics uh jason in very much a workshop format uh engaging people so hope to see you all there um but these issues and attend thank you so much for putting this virtual event together uh and your team because these issues are so critically important to our industry and um hopefully we can see some of you live at auvsi in florida thanks don yeah no shameless plugs are always allowed absolutely 100 there's no

**[41:08]** rules against this this is this is our panel we could say what we want but we provided so much value today you know what and but before we continue tin what are your thoughts how far is too far you are literally this summit is your summit your company's putting this on how far is too far right so I mean of course now gemma is the person who has most closely you know worked in these situations and she's aware of I know what works what doesn't work and what could cause or reversal the situation by by using robots or by using autonomy and she is right but at the same time as a technologist I would like to believe that you know we need to kind of dig deeper identify what the problems are and you know work on solving one at a time and keep improving the systems till we get the desired results like for example I know 10 years ago who would have thought that autonomous cars would be driving on our streets right and um I mean I remember 10 years ago we used to have fierce

**[42:10]** debates of similar nature that autonomous cars are never they're never going to happen and we need humans driving the cars and taking decisions on know what to do because if somebody loses life because of an autonomous car then who is to be blamed and uh no what was the algorithm that was taking the decision to kind of do whatever the car did and so on right and now uh ten years later I think uh everybody agrees that autonomous cars are going to happen it's just a matter of whether it will take another couple of years or ten more years but but it's it's going to happen right and we already now have cars that have driven know millions of miles and they have proven that they are safer than human drivers uh right so on on basically as again know as a technologist while I appreciate all the concerns that gemma has and uh of course those should be addressed and we have to be mindful and we have to design systems to uh know kind of address those uh but I I still believe that no

**[43:12]** if we take up one problem at a time and we keep investing in building better and better systems we will be able to design systems that can kind of address and solve those problems it might take us a long time it is not the low hanging fruit that we discussed earlier know in our conversation it's it's a far higher or for difficult fruit to uh look get hold of but but I think it's it's not impossible and maybe 10 years later we will be having a very different conversation 100 I think this entire industry and I see the live clock we're clicking down I think we're almost at 45 minutes right now this has been amazing and I highly recommend like I said follow follow follow this panel we're all doing spectacularly amazing things in this industry and we are just here to support the industry overall uh check out what we're doing this has been a great panel we kind of wandered away from kova 19 but kobe 19 has definitely been the impetus around a lot of what's happening in this industry it has truly pushed this industry exponential it was

**[44:14]** heading there anyway it was just a little bit of nudge that we needed to sort of get us there so I want to thank the panelists thank these ceos thank you so much this has been amazing so much fun and if we have any other questions we could take them otherwise thank you so much for coming to this fireside chat my name is jason san souci I'm also the host of the do you know drones podcast another shameless plug feel free to subscribe to that on youtube and wherever you listen to us in your podcast and we hope you have a great rest of this session thanks guys thank you thanks you