DEVIN GLADDEN: Hello, everyone. Welcome to the first episode of transportation tech talk. I'm very happy and excited to be able to share some really interesting insights with you all on a variety of tech topics that my cohost and I believe will shape the future direction of mobility. And we're excited to share
our perspective with you and to get more thoughts out there and to get the conversation going as mobility is a topic that impacts us all. We all have many thoughts around how do we get from Point A to Point B, and how do we get goods from Point A to Point B as well. And so, we all have to agree that transportation plays a really important role in making sure that we see the movement of people and goods across this country, across the globe, and we understand the huge social, economic, political impact that transportation and infrastructure and the investments and the politics and the policy, and I can't wait to dig in to all of that. One quite note before we start: All of the opinions represented here are from the hosts. Although we work for and have worked for a number of institutions in the transportation space, today we're representing ourselves and sharing a little bit around how we think these topics and technologies might unfold in the coming years. And so with that, I'm Devin Gladden. I should have started with that when I started [laughter], but
I'm a technology and public purpose fellow at the Harvard Kennedy School in the Belfer Center, which is a think tank where, for this particular fellowship, seven folks have come together for the full academic year on a range of technology topics, really to push the envelope, to ask in-depth questions, to make deeper connections within the industry, and to really help to move the needle to ensure that when technologies are deployed that we maximize benefits and minimize the harms. And I'm excited to be joined by my two cohosts who are illustrious in their careers and have done and will continue to do some really incredible things. And I look forward to hearing their insights. And so with that, I'll turn it over to Agata.
AGATA CIESIELSKI: Hello, folks. My name is Agata. I am a presidential innovation fellow, and we're a group of movers and shakers brought in to bring in innovation into the government. However, that's not who I was in my past life. This is just who I am, part of who I am now. I am actually a trained roboticist and AI researcher. I am on leave actually
from Johns Hopkins University to help out the US Department of Transportation bring AI into our next generation of infrastructure for transportation. So with that said, why am I here? [laughter] And to be honest, transportation is just one of the most exciting things. It doesn't sound exciting maybe when you think about it at first glance, but then you start to realize transportation is vehicles, it's quad copters, it's robots, it's how we get those packages and our gifts. It's what makes our world run. It's so interesting,
diverse, with so many different problems to solve to make our space a better place. So with that, I'll hand it over to Kamya. KAMYA JAGADISH: Thank you. Hi, all. I'm Kamya. And similar to Agata, I am also a presidential innovation fellow also serving the Department of Transportation. In my former life,
before being a presidential innovation fellow, I was a data scientist in product management at various technology companies, including Lime, the electric scooter company. So transportation has been very interesting to me for a long time. And I think part of the reason I was drawn to transportation earlier on in life is just from some of the personal stories that I had in terms of the independence that I was able to feel when I could get around the city by myself as a teenager, or when I would be in a new country and I was able to feel safe getting somewhere, even though I didn't really understand the terrain. And I think that sort of independence
as an individual is something that everyone should be able to have whenever they're getting anywhere, and is a huge reason why I'm excited to be working in the transportation field. And will hopefully continue to do so for a while. DEVIN GLADDEN: Great. I'm so excited to be joined by Agata and Kamya. So today we're going to start our talk with mobility on demand. It's a relatively new technology process intervention in the tech transportation space. So Kamya, what is it? Talk to us a little bit about the features of it.
KAMYA JAGADISH: Absolutely. So mobility on demand, or MoD is the acronym that you might see floating around the Web, definitely has a few different definitions, depending on which research institute you might refer to. But overall, I think what mobility on demand really is, is a model of transportation that leverages innovative technologies to actually help people get from Point A to Point B, regardless of it being a route that might be low density or at a time that's sort of off-peak. And so, I think it's this really dynamic way for people and for goods to actually seamlessly get where they need to go. Now, isn't that sort of like the vision of transportation in general? I would say yes. So personally, I think that mobility on demand is really just the future of transportation.
Public transit can't serve all needs because, as I said, there are these sort of like edge cases of off-peak or low density routes, and we still want to be able to serve those people and help them actually end up at their destination of desire. And so, that's where mobility on demand can really come in to play, really being able to call a car at that first last-mile gap that you need to be able to close, or see a bike share station there that'll help you get there. And so, I think that's sort of the vision of mobility on demand, is really thinking about how to use innovative technologies and solutions to sort of complete that whole trip for someone. DEVIN GLADDEN: It's interesting now that we're focused on trips. Understanding that for a lot of people there might be various modes of transportation included in one trip. I think about before the pandemic, when I was traveling a lot more,
sometimes you might fly into a city and then if you don't have a rental car, then you go on an app, you get a ride. Then you get to your destination and then maybe if you're in a dense city, you might consider taking a bike to do a day trip or a rail system. And so, it's thing, I'm really curious, in all those different modes, we've got so many different stakeholders, so many different agencies involved. How is that coordination working? KAMYA JAGADISH: That is a great question. So I think there's actually two things there. One is the technology integration and one is the sort of like stakeholder partnership integration.
And so, the technology integration is trying to think through, how can you actually, as a user of transit, take that plane and that train and the bike and the bus and actually be able to seamlessly be able to understand which mode to use when and maybe have an integrated payment system as well. So I'm always just sort of able to use the same fare pass essentially. So that's one piece of it. But I think the piece you're asking about is what I find super interesting, and it's actually the partnership piece. Transportation in general engages public and private partners, but I think mobility on demand especially engages private partners because a lot of these solutions, like TNCs like Uber and Lyft, kind of serving potentially that last mile in public transit routes, or scooter or bike share companies, a lot of times those are those private companies and they're having to build in relationships with the public sector. But then on top of that, you layer the fact you have local transit at the city level, and then you also have the federal level. And so, trying to think through what role should the federal government play, what role should local play is pretty interesting.
And so, in harkening back to my previous response, I think federal and local need to care, right? This is a way to actually get people places and so they want to be bought in to this idea. And I think if you think about what's actually like a DoT's mission, it's to achieve this vision of safe, efficient and modern transportation, and that to me really is mobility on demand. But when you think about what the role should really be, at the federal level I think they're really well suited to help guide local cities actually implement these technologies. There are so many different technologies coming out of so many different fields from different private companies, to academic institutions, and they might be seeing different cities implementing different solutions, and it's really hard to navigate so many different technologies and solutions and partnership models. And the federal government can really be of aid in here in just sort of helping these local transit agencies, which are often underresourced, actually navigate the sort of decision process and implementation and adoption process.
So I really think that's one area that federal government can be super helpful in achieving these sort of mobility on demand dreams that we have. One thing that I think at the local level they're really well suited to do is ensure there is an emphasis on equity and actually meeting needs of underserved communities, because it's sometimes harder to actually see if that's happening at the federal level. So as an example, the Federal Transit Authority, or FTA, had a mobility on demand grant that they awarded a few years, I believe in 2016, called the Mobility on Demand Sandbox. And they were awarded a few different grants. Los Angeles and Puget Sound got two of the larger grant amounts. And they were basically implementing some creative mobility on demand solutions.
Eno Transportation did sort of an analysis to understand how well it worked, and it seems like showing that technological integration of actually combining private and public transportation methods technologically, it can work. Which is pretty exciting to see. The partnership also can work; there's obviously improvements to be made. But one of the things that is still sort of an area to be worked on is understanding how to make sure that the people who need these sort of extra modes of transportation most – maybe they're lower income or they live sort of on the fringes of the city, for example, and don't have ways of getting there, or maybe they're disabled and they live in a rural community, and so it's harder to be able to take just normal public transportation or they can't drive their own car – how are we making sure that we're also using this to meet those needs? That's something that we didn't see as much of, or I didn't see as much of in the FTA awards, and so I think that's one area that local governments can really play a big role in shaping sort of how we're ensuring we address equity standards while we're implementing mobility on demand solutions. DEVIN GLADDEN: That's an awesome understanding of how complex the ecosystem is and where we might be going. And in some ways, to your point about the federal government, because we have so many different governmental jurisdictions that are going to be involved with managing and overseeing the development and deployment of these programs, the federal government does have a unique role to be able to see everything, but you can't act on everything, but you're funding a whole lot and you can certainly look at how you can get certain objectives achieved at a local level through funding. But a lot of it will be up to the local decision-makers, those who are using the system day in and day out.
Your point about the federal government being able to provide guidance will be really critical because the federal government's looking across and they're coming in contact with all these cities and states that are doing all these different things and they can apply best practices and to help avoid some of the worst practices, too. I agree, I think there's an important role for the federal government there. KAMYA JAGADISH: One more thing on that note, is I think that there's the opportunity also just to share knowledge between cities. It's this transferability of what we're learning, right? Mobility on demand is, in my opinion, very likely going to happen. It's already happening in
a lot of cities. And we want to do it well and we want to do it efficiently and effectively. And there's so many cities across the country that are already trying to do these things, and will continue to. And so, let's not waste time and waste resources by trying the same solution everywhere. Let's really transfer that knowledge. And I think, again, federal government can help in sort of shining that light and trying to make sure that there's an ability to crosspollinate and crosscommunicate in a way that will hopefully be really effective.
This can obviously happen grassroots up with cities just naturally talking to each other. But it can be so much more centralized and sort of structured and efficient if it's also structured top-down. DEVIN GLADDEN: Agreed. I'm thinking, too, this is an interesting area where data will be very important because you're working with so many different stakeholders and institutions who are all trying to implement their piece of the journey and trips. And so, I'm really curious,
how does data play in to this? And Agata, this might be an area for your expertise as well. AGATA CIESIELSKI: Sure. So for starters, Kamya, thanks for introducing us to that. I cannot say how excited I am to honestly not have five gazillion apps just to get 30 minutes away from my house. But that being said, data is everything. As you all know, we are in the digital age and it is an amazing and beautiful thing. Data is really, it can be scary, but it can also make our world
totally and amazingly efficient. And we can really kind of get back to being human and having our technology really work for us instead of us working for our technology. So data has actually been identified by the federal government as a strategic asset, which means that in the federal space we're going to be seeing a lot more emphasis on data. And not only is there going to be a lot more data, we're trying to facilitate this culture that values data and promotes its public use. So we're going to start to see data be a lot more open source. We can already see it with the Department of Transportation; we have transportation.data.gov, which is connected into
data.gov, where you can get tons of public data sets. And when we have data, we can start to have this amazing grassroots approach to innovation. We can have not only companies who are developing these newest, coolest hovercrafts that will get integrated into mobility on demand, or solve our annoying problems, like why are there five gazillion scooters when I've got three trucks over there? And it's going to help just make our world feel a lot more seamless, and we'll be able to have everybody across the US and abroad innovating for our own spaces which is an amazing and powerful thing. And we're going to be able to do that through data. KAMYA JAGADISH: I fully support everything Agata's saying. I totally agree, data is everything. It's the way to power the visions that we want to have. And so, we want
to have, or I want to have this mobility on demand vision. It will solve a lot of people's problems, but from an equity standpoint, just from an efficiency standing, a safety standpoint. And I think the technology that'll actually get us to that point, it's not going to be, in my opinion, it's not like autonomous vehicles are going to help us complete our mobility on demand vision. I think that the technology is not the flashiest thing; I think
the technology is going to be data infrastructure, how are we actually taking in the data from the private partners, the public partners, of which there are so many, and integrating all of that data and processing that data. And then putting that business logic layer on top and actually making decisions with that data and making it palatable in a user interface for people. I think those are the actual technologies that are going to get us to this next version of our transportation future.
DEVIN GLADDEN: You all are making some interesting arguments, particularly in favor of more data extraction from the public and more open and transparency around it. But I also think about the real concerns, particularly around privacy that people have around data and the desire to be able to, particularly in the transportation sector, to move freely and to be able to have, if you so choose, to have a certain level of anonymity. But I wonder, as we move to more of a mobility on demand era, where all of this will be coalescing around one app or one approach to all of your trips, I'm really wondering, how are you all thinking around privacy and ensuring that, as consumers give more data up to bring the vision of mobility of on demand to life, how do you ensure that they're still able to feel protected in the ways that they like? AGATA CIESIELSKI: I'll jump in on this. For starters, we are all very concerned and value the privacy of our data. But that being said, we have to remember that culture and ethics and all of these litigations, they're all fluid. They all
change from day to day. What is okay today might not be okay tomorrow. And vice versa. So what we're really, really trying to do is to make sure that we have this infrastructure that allows for the changes. So for example, I think if we decide that we don't want cameras to be helping guide some of our traffic systems in the cities, there are so many different ways to solve these problems, which is why transportation is so exciting. But then what we can do is we make
sure that our systems are flexible enough so that we can take these sensors down and make sure that we're– again, the goal is to make our technology work for us and not us for the technology. We want to be making sure that this is bringing value and goodness in our lives and it's not going to be if we feel like we're being surveilled by our next-door neighbor or Big Brother. So with that, I'll pass it to Kamya because I'm sure she's got some thoughts. KAMYA JAGADISH: Absolutely. I think that what you said makes a lot of sense to me. And I really love the point about actually also showing people what we're doing with the data.
I think that's key in terms of transparency, but also just in terms of showing the value. I don't think we should be collecting data that we're not using and just sitting on it because, of course, it's going to foster feelings of mistrust and also just confusion. And so, I think it's important to also ensure that if we are using data for putting sensors or cameras out there, making it really clear that we are using that data and how.
And then, as you said, if it's not working or if it's wrong, or if we decide it's ethically bad to do, we should have a way of reversing it and taking it down in a quick manner. And so I think that's a really important point that you highlighted. DEVIN GLADDEN: As I think about another big technology being used in transportation space, artificial intelligence, it is so data-intensive. Partially because you've got to train the models in order to make predictions and to give you insight. It's got to have something that
it's rooted in and information. And so, there's a lot of data training that happens. And so, Agata, I know this is your area of expertise, so I'm really curious, how do you see AI playing out in the transportation space over the coming years? And what role will data play in that? AGATA CIESIELSKI: Well, as you just mentioned, AI is driven by data. However, like I said, and I keep harping on this, this is going to be innovation that is for us as the American people.
And so, I think a lot of the things that we're going to start seeing are actually going to be a lot more invisible than you would expect. We're going to start to see, for example, AI being used to do things like inspect our roads so that we don't have to be hitting five gazillion potholes on our way home. It's already being integrated in things like traffic management systems so that we can have cleaner, not to mention more flowed traffic so that you're not stopping and going 20 gazillion times on the way home. And then in other places, you'll also see it in safety, so we can do predictions on where accidents are going to happen based on weather conditions, and then we can quickly deploy emergency services. So I think there's so much space for AI in transportation in places that you wouldn't even really think of as a daily consumer, just kind of going through your world. But
it's going to happen and that's where we're trying to take AI in transportation next. DEVIN GLADDEN: I know a part of your work has been a federal data strategic plan, and I'm really curious, how is that fitting in to your work? And where transportation fall in all the different aspects that we're seeing data apply? AGATA CIESIELSKI: So as I mentioned before, a few years ago there was a federal strategic data plan that came out, a mandatorium[?] that came out which said that data is an asset. And within that strategic plan, it's actually highlighted that we want to promote open data and sharing and reusability. Because frankly, it's early expensive
to generate data, to generate good, clean data that can actually create the solutions that we want to be creating. So the strategic value is just infinite. And how it's going to be used, it's going to be used in so many different ways. You can imagine, for example, having weather data being used along with transportation data, being used with our schedules to help promote more efficient transportation. And this is going to happen
through these more co-innovative solutions that are going to be happening across places that you might not even expect. Which I think is vastly, vastly exciting. Did I get what you were getting at? Do you want me to dive in to it a little bit more? DEVIN GLADDEN: I mean, listen, we've got time, so yes, let's make it for data. [laughter] KAMYA JAGADISH: Agata, I'm curious your thoughts, in the next decade or so, how do you imagine this innovation and obviously with data as a huge piece of it how do you imagine it's going to drive the deployment of new on-road technologies? AGATA CIESIELSKI: Great question. So I think we've seen that this space is changing so quickly. It was only, it's only really been ten years since we've seen the first major
neural nets show some great promise, and now we're seeing it almost everywhere. So that being said, it's still a really expensive and difficult technology to deploy. But I think that when we start to have innovations in the methodologies of scaling, we're going to really start to see crazy and fun and new things be deployed. And that's going to come from, there's going to be advances in computing. There's going to be decrease in costs. We can have small sensors and computers deployed to advance and put AI in spaces that we didn't think was possible before.
And with all that investment into this new digital infrastructure, we're going to be able to have this space that allows innovation from everywhere. And to add to that, we've now had ten years of a workforce that has been building with AI knowledge. And it's only going to get faster. I think we're going to start to see amazing solutions that
will make our future just look like something out of sci-fi. I'm picturing personally the Jetsons. I want to see to see a Rosie Robot, and I want to have my little flying car. Personally, I think that would be amazing. But really, the limits are boundless.
DEVIN GLADDEN: That's a really excellent point, particularly when you look at how complex the challenges that we are dealing with in society and what we're hoping to achieve through our new vision from mobility. And one of those things I'm most interested in and honestly this is why I'm focused on the transportation space, and I've dedicated my career to understanding how can we reduce climate emission from the transportation sector? I think this will be, to your point, Agata, around the breakthroughs that we're going to see. I think in the coming decade, climate change will be one of those issues where we'll see new insights gained from AI applications and new solutions because we might learn a couple things from our robot friends about what it means to be human and to protect our planet. And I'm excited about that because I think it will bring about a better future because we'll be able to actually solve the problems that we have; climate change in particular.
AGATA CIESIELSKI: Well, to jump on that, it's actually a priority for this administration right now, which has been pretty exciting. We actually have the climate change center that just has been launched out of the US DoT. And I think you're going to see some really exciting work. Personally what I want to see, again, just my own bias, I want to see places like NOAA working together with DoT, working together with Department of Energy and EPA to test out all the innovations that are coming out with clean energy and more efficient driving. And then we can see
how they impact our climate models and make sure that they're really being deployed in a safe way. And like I said, I love this idea of co-innovation and I want us all to work together to get this utopia that we all, come on, let's just admit it, we all want it! [laughter] So I'm going to pass it back to you, Devin. DEVIN GLADDEN: I actually, before we move to the next topic, I actually have one more question because you brought up a really good point about the co-innovation, and that makes me think about the consumers. As we all appreciate in the transportation space, if
you build it and they don't come, you don't have anything. [laughter] So I'm really curious how you are thinking around what considerations consumers should be thinking around as they experience and test out and adopt these new and emerging technologies. What should they be thinking about? AGATA CIESIELSKI: Let's start out by saying that we should all choose the technology that we are most comfortable with. But all of that being said, I think a lot of these things, as I mentioned before, they're going to be pretty invisible. We want to add AI to make our everyday
lives just feel more seamless. So I think as a consumer, I think we're going to start to see things jump in that I hope that we're all looking for. Some of these things are perhaps night vision to have AI systems that work better at night, which right now, to be honest, is a struggle. To have things like pedestrian detection, to make sure that our vehicles are treating the pedestrians that are walking around safely. We can have things like traffic sign recognition. So essentially I think there's going to be a lot of capabilities that are coming in to the personal space, as well as those invisible things that you're seeing. And then finally, I think
we're going to start to see technology that can come in to the areas that we don't expect it. Which I'm really excited to see because I'm here in West Virginia, which is a far cry from where a lot of these technologies are developed, and I want to be able to see some of these technologies hit this area where people really aren't really thinking about us all the time, to be frank. But yes, as a consumer, I think you're going to start to see things come in to make your life a lot safer, a lot more seamless. But again, choose the technologies that you feel comfortable with. DEVIN GLADDEN: I think that a really great point about choosing appropriate technologies. I think that's a great segue to our next topic on autonomous vehicles because my work has shown me that the deployment of these vehicles will very much depend on the use case. And as you noted, being in a rural community versus an urban community, needing an AV for a last-mile ride from a train station, to delivering lunch, these are all different use cases that I think as we see the industry continue to mature and develop, we're seeing more of that being built out. And that
is leading to some interesting partnerships and acquisitions. I think about just today, it was announced that Lyft's AV division will be acquired by Toyota. I think some people will find that information a little bit surprising because Lyft, and Uber before it sold its unit to Aurora recently, they had really been real champions in the space of the technology because there had been a lot of discussion around, will these vehicles be able to reduce the labor cost of having a driver, so much so that you could end up with a dollar-per-mile-per-ride, which opens a whole new equity conversation when you start to make transportation options more affordable to people, particularly if they're in an area or community that has poor transportation infrastructure and a car can, when you're able to figure out what's the best use case.
AGATA CIESIELSKI: I want to flip the last question that you asked me straight back at you. So what do you think? AVs, honestly, when people think of AI in transportation, they think autonomous vehicles. So what do you think? Which are the consumer applications that are likely for AVs in the coming years? What do you think? DEVIN GLADDEN: To start, I hope that one key takeaway from this talk that people take is that AI will not only be applied in AVs, and actually the more transformational applications, which, Agata, I agree which will be traffic management, are much bigger applications that the public should be following, just as much as they're following Teslas and the comings and goings of AV acquisitions in the industry.
But to your question more pointedly about where we see these things going in the applications, I actually think in the short term, we'll likely see long-distance and short-distance deliveries be sort of the first deployments. Because one, I just think you actually, in the trucking industry in particular, you're dealing with severe labor and driver shortages. And some people have also been looking at the fact that because of the pandemic, that we've really given rise to the delivery economy; everything can be delivered. You can even get a car delivered now. [laughter] Which I find hilarious. But in some ways, it's this need to be able to have goods really delivered all places, at any time of the day. You have to be able to facilitate that. And I think that's a really interesting application for AVs.
I do think though the larger question around whether or not we're going to see delivery bots everywhere will really be how drivers really acclimate to this new road user, where in some ways– you know, we've never really dealt with robot drivers before. Very few places in society. But it will raise really interesting questions because in some ways looking at new road users and understanding that humans have developed all sorts of unwritten/written rules around the road and you've got actual legal requirements, federal, state and local. I mean, this makes me think about our earlier conversation about mobility on demand where you've got so many different organizations layered all in trying to make the whole system safe. And that's exactly what you have when you look at regulations between the driver, which most people get their driver's license from their state, and your state regulates you as a driver, but in AVs, who's the driver? Is it the robot? Is it the automaker? Who is it? So there are still lingering questions in the regulatory space that I think will need to be answered that will support whether or not we see greater adoption and use of these vehicles.
KAMYA JAGADISH: I think that really touches on a lot of the things that I am most curious about when I think about AVs and delivery vehicles. And so, I'm so curious, in your perspective, what do you see as the steps we need to take to be able to start addressing those kinds of concerns of those other road users? And how do we even try to start mitigating between industry and regulators? I realize that there's no perfect answer to this, but I'm just curious to know what your take on that is. DEVIN GLADDEN: Thank you for the question because this actually touches on a really important aspect of my research this year, which focused on what I'm calling four elements of public acceptance. So these are four critical areas that any government, both state, federal, everybody involved with regulating AVs, but we all need to be considering and actively address because these are items that the public needs if you're using the roads just like anybody else, these are four elements that people really do need answered before they feel comfortable driving alongside a robot. And so those elements are: The first one is that AVs demonstrate safe and predictable behavior. So this gets to the first point that I was making about the rules and regulations, unspoken and the formal regulations, but really ensuring that AVs can follow those rules and that they can be a reliable, trustworthy road user. Until folks understand that AVs have those capabilities and are trustworthy,
it's going to be very unlikely that they're going to feel comfortable driving alongside a vehicle. And that kind of tension and anxiety can create stress, which is– I have to remind people sometimes, but driving is still one of the deadliest forms of transportation. A whole lot of people die every year driving on roads, and so you have to imagine if people are already in a situation that makes them a little precarious or tense on a highway, and then now they realize, Okay, there's actually not a human diving that vehicle and now I'm a little concerned. How do I communicate with it? Do I just go around it?
People's behavior might start changing. And our research showed that people have questions, and there's this preexisting anxiety that regulators really need to address. The second key issue is that the technology actually works. I think about Agata's point earlier about pedestrian detection. We've seen studies that show that these technologies don't
operate as they are designed. And so, I'm a '90s kid; I remember having to blow into a Super Nintendo cartridge to get it to work. I'm not sure people are willing to troubleshoot on the road when they've got a car full of people and they're concerned about this robot next to them. And that will be a key industry perspective because they're going to be responsible; it's their technology on the roads. And regulators need to be able to strike the balance to be able to kick the wheels on it, ask all sorts of interesting questions about the software – how does it make decisions? How are you defining redundancy in the hardware so that a car just doesn't stall out? That will be a key consideration.
The third element is that road users really have a key understanding around liability for accidents. And I think because we're still in this nebulous pilot phase where there's a whole lot of pilots around the country in different parts, but some people have not seen them, and so there's still an open question around what happens with liability. And I'll be honest, I think the insurance industry has been kind of tiptoeing around this, state regulators, where insurance is regulated in this country, they've been tiptoeing around this. I actually have a different approach because I do believe that if industry is willing to put all of these investment dollars into these vehicles to bring them to life, I think that means that they should take on the liability, at least for a short amount of time. So in some states,
I think that for a set amount of time they should consider placing liability solely on the entities testing the vehicles. I think that would give road users a sense of confidence, at least in this interim phase, when our public roads are being used as an experiment, and to some extent we're seeing the impact of that with, unrelated to my research, but we're seeing that with ADAS technologies. I mean, Tesla crashes generate a whole lot of media and public interest, and it's one of those areas that I think about where, if we can't manage ADAS right and we can't show people that we're going to be responsible developers and assessors of this technology, how can we ever believe they will give us their full trust in a fully autonomous vehicle. We're not there yet.
And so, I think liability will be a key consideration. And hey, it feels like every other day the industry is willing to throw around another multibillion-dollar evaluation, so let's think about that in the liability sense and let's have this public testing really create a safety net so that we can get the full benefits of it and minimize the harms. And then the last element is making AVs recognizable instantly. I don't know if it's a marquee on top of the vehicle like you have with cabs. Is it some sticker, like when you have student drivers, they've got a little bumper sticker. But something that just immediately gives
folks, other road users a signal that the car in operation or the vehicle, the semi truck, is not being operated by a human, and it may not have a human inside of it. I've had conversations with professionals and I've seen inside of AV shuttles. Never driven alongside one. And it's interesting, I've talked to folks who have deep, deep AI expertise and they still say, "I'm not sure what I will do when I pull up next to an AV for the first time. I'm not sure how I'm going to react. I'm not sure, am I going to have different expectations?" And I think all road users are going to have [laughter] that first kind of "huh" moment. And I think that's where I want to be able to step in with some advice, some information, giving people a clear sense of what the capabilities and limitations are of the technology. This is getting into Consumer Education 2.0 for me because we talk a lot
about– in the industry there's a lot of talk around, okay, well, if we just educate consumers, that will help. But I'm always like, what are you talking to them about? And I think this is kind of a key area to start with, like how to drive alongside an AV. And I wonder, these testing scenarios, there's a lot of openness there for people. And I think we can do some good work, but it also opens the door for a lot of damage if we don't manage it well.
AGATA CIESIELSKI: So Devin, I'll chime in with a funny story. When I was in graduate school at the University of Pennsylvania, we had access to– we were fortunate enough to participate in the DARPA grand challenge, which, for those of you who don't know, is a challenge that was hosted close to ten years ago to develop a fully autonomous vehicle that could go from Point A to Point B, to nearly around 100 miles fully autonomously. So after the program finished, the challenge finished, we had access to this vehicle. And
we were doing some research at the GRASP Lab at Penn, and we got some city permits to go in downtown Philadelphia and we were testing actually how pedestrians were going to be interacting with our autonomous vehicles. And so, we were so ready to have people looking backward and for our algorithms to not work properly, and much to our sadness and dismay, no one even noticed we were out there. [laughter] We were so disappointed. But that being said, I think you raise a really great point. It's not about finding the right solutions always; it's about how you present the solutions.
And I don't know what that's going to be. When you first said that, I imagined vehicles with this crazy bubble [laughter] or like big bumper car style. But of course, I think we're going to have to think of the best way to do this and hopefully pull from innovators from industry and academia to come up with those best ways of what should these vehicles look like when they're actually deployed. DEVIN GLADDEN: Yeah, and you make a good point about vehicle design because I'm expecting– you might see all sorts of new vehicle designs come out, especially if– I think automakers are really keen on trying to promote alternative seating arrangements and activities in vehicles.
So if you think about now, in your hour-and-a-half long commute home from work, maybe you'll be able to sleep now. There are some interesting safety questions that you ask now, like, is it safe to have people lay down in cars? But this is where, to me this is where– actually vehicle design is a key area where we should have greater collaboration between the federal government and industry. And we will need that greater collaboration just to ensure that everything we're putting on the road is safe. And that will be a big question, a big public conversation constantly around what is safe? Is it unsafe? Who is it safe for? Why is it safe for them and not for others? We see some of this already when we are actively engaged with auto regulations. But I think we're going to see more of a public conversation around it just because of the real hype and mystery around these vehicles and what people think that they will be able to achieve and what they want them to be able to achieve for them to be real in real life.
KAMYA JAGADISH: Yeah, absolutely. I think that you both are making really great points around the public trust aspect of it, which I think just can't be underemphasized. Sitting between of you in this Zoom room, I have not ever been in an autonomous vehicle before, and this is something that I know Agata has really talked about passionately before. But there are so many people who are in the rooms making decisions about autonomous vehicles and who are in the transportation space, such as myself, who know some stuff about autonomous vehicles, but I've never actually sat in one. And so, definitely it's going to be harder for even myself who, I would say I'm data-driven, science-driven, et cetera, to really feel fully trustworthy of the decisions that are coming out of the industry/government in this regard. So I think that, yeah, you make a really good point in terms of making sure that we're really emphasizing the fact that we need to be talking with the people and having this public conversation and bringing people into it.
DEVIN GLADDEN: Kamya, that's an excellent point. And I think that brings us to our final segment for this episode, which I'm very excited about because it gives the audience more understanding into how we intend to use our fellowship experiences and the insight we've gained. And for me, I think the biggest insight I've gained about this sector from my fellowship, actually we touched on it earlier, but it's about how energy and data-intensive AI technologies are. And in some ways, I think as we, as humans, become better stewards of the planet and we make larger, more inclusive goals for the transportation system in our country and globally, I really do think we will have to really, really investigate and manage the expectations around energy in particular and data because these systems are very, very intensive. And I do think because they require just so much, there will always be a public conversation about, are the tradeoffs worth it.
And I think it will be up to people like us, who are in the industry, who are using our careers to impact the deployment and trajectory of these technologies, to make sure that we're being honest with folks and saying the truth, being truthtellers. Because I'll be honest, I think the public is skeptical of industry and what its motivations are. And to be honest, people are also skeptical of the government at any level, and what's its motivation and why does it care and what actions will it take? And I think the more we as professionals can help to steer that conversation, to guide and make sure that the correct information is out there, that we're asking the right questions, that we're thinking about the needs for everyone, I think we'll be in a really great position to do some great work. And I think this will be the first of many conversations between us on these topics. AGATA CIESIELSKI: I suppose I'll jump in after you. So what have I gained out of my fellowship?
I will say one of the most surprising things for me is learning how much I actually like civic tech. It is one of the most powerful and exciting concepts, to be able to come up with ideas and solutions that actually impact me and everyone around me in an everyday basis. And I think the most amazing and pleasant surprise is that as I'm navigating through these spaces and working through these different agencies, there is so much excitement internally as well.
And I think there's a real desire to drive some change and to really bring innovation into our spaces, to do it from a grassroots or a ground-up way of thinking. And so, whether or not I decide to stay in the federal government, I think one of the most important lessons that I've learned is, hey, I love working on big, really difficult problems. And I just love making an impact. And I hope that one of these days we'll be able to see the fruition, and all come together and look around and say, hey, look, we built this together.
KAMYA JAGADISH: Following on to Agata's big ideas comment, I totally second that this is such an interesting space to be in because it's so complicated. It touches everybody as humans. So many different industries and public sector at every level and academia, et cetera, are involved. And so it's really complicated. And I think that's what makes it super exciting.
And something that I maybe naively have thought in my past when I looked to a problem that involves public and private sector, I haven't understood why it's so complicated. It's so easy to be a human just consumer of transportation, for example, and just say, Well, why can't we just bring out autonomous vehicles and make that be more ubiquitous and, I don't know, make it safe in the next ten years? Maybe that just seems simple to some people because you don't have all of the knowledge about how complicated things like this really are and all these partnership models are. And so I think that's something that I've really been learning a lot over the past number of years, but especially working in the federal government and understanding the federal perspective on how these things work. It's just been super interesting.
And so, one of the things that I just feel really strongly opinionated about is just the idea that you always need to keep trying and iterating on your approaches, not just to technology development. I'm coming from a private tech background so I know that with technology development you're always talking about iteration and learning and in Av testing and retrying again. But you need to do that also just at a larger level with how you're thinking about a partnership model, how you're thinking about a large overall strategy for a department. It's not just about the technical product. And I think that's something that that both Devin and Agata have sort of mentioned in their responses to other questions earlier today, whether it's about testing out autonomous vehicles and seeing what types of solutions actually work, or whether it was Agata who I think said this really nice phrase about innovation methodologies to scaling, which is another thing – how are we scaling these technologies? We need to think innovatively. We have to have co-innovation across cities and see what works and what doesn't. And learn from it.
And I think that applies to mobility on demand, obviously, as well. Really just thinking about how do we actually want public institutions to partner with private institutions. We're not going to get it right; we need to keep trying different procurement methodologies. And I think it's sort the sandbox approach that we want to take
to everything we're doing in these big, complex problems, not just the technology itself. DEVIN GLADDEN: I agree, Kamya. And I think that's the one takeaway that I think we can all agree on, is that human use and adoption of the technology will be critical here. And we have an important role to play in guiding it. And so, I'm excited. And thank you so much for this great talk. I'm so happy that we were able to just share
some insight on some other technology topics that I think we'll be certainly hearing more about in the coming years. And people will be seeing and experiencing on the roads. And that's when the rubber will hit and we'll figure it out along the way. So with that, thank you so much, everyone. And until next time [waves]. KAMYA JAGADISH: Thank you.
END
2021-05-21