Autonomous Vehicles are Ready to Disrupt Society, Business, and You

BRIAN KENNY: At the 1939 World’s Fair in New York City General Motors unveiled Futurama, an exhibit spanning an entire acre that featured a model of what US roadways would look like 20 years into the future. Networks of streamlined motorways wound through a landscape of half a million buildings, a million trees, and 50,000 miniature cars traveling on a 14-lane, multi-speed highway. It struck a chord at a time when the country was just beginning to grapple with traffic congestion. In what might have been the boldest prediction of all, Futurama depicted a future where self-driving cars would communicate directly with the road moving passengers safely and swiftly to their destination. It seemed like science fiction, but by 1958 GM made this concept a reality with one of the first full-sized self-driving vehicles. 50 years later, self-driving cars are still a work in progress. Today on Cold Call, we’re doubling down to look at two cases that each look at the future for autonomous vehicles. I’m your host, Brian Kenny, and you’re listening to Cold Call, recorded in Klarman Hall Studio at Harvard Business School. Joining me in studio today is professor Bill Kerr to discuss his case entitled, “Autonomous Vehicles: The Rubber Hits the Road… but When?” Also in studio is professor Elie Ofek to discuss his case entitled, “Autonomous Vehicles: Smooth or Bumpy Ride Ahead?” Thanks both for joining me today.

BILL KERR: Thank you.

ELIE OFEK: My pleasure.

BRIAN KENNY: It’s great to have both of you here. I found out about Bill’s case first and reached out to him to do that. Elie, you and I were going to do a completely different case, but somebody brought to my attention that you had also written a case on autonomous vehicles. The two cases are really complementary, and I think this will be a really rich discussion about a topic that is certainly something that’s been in the headlines a lot, and I think it’s one of these things… I know, speaking for me personally, I can’t wait to get into a car and just open up my newspaper and let the car take me where I’m going. For people who are listening to this while they’re driving in their cars, keep your hands on the wheel. It probably doesn’t steer itself yet, but it could very soon, right? We hope. Anyway. I’m going to start out, I’ll start with you, Elie, but I’m going to ask both of you the same question. What is a world that has fully transitioned to autonomous vehicles look like, and how is it different than today?

ELIE OFEK: Great question. When I teach this case, I usually ask some students to just close their eyes and have exactly that vision and then describe that vision with their eyes closed. You could have various ways to think about it, and every person has their own vision of what that looks like since we don’t have it in front of us just yet. But a world like that, if you could picture a lot of room buzz around but not going in a random direction but rather-

BRIAN KENNY: Not bumping into the couch.

ELIE OFEK: Exactly, not bumping into the couch but rather being a little more functional about it, efficient about it, effective about it, a world where, as you said, maybe the driver is no longer a driver, just a passenger, and you can do a lot of stuff in the car, a scenario where there’s far fewer vehicles on the road because we can utilize those vehicles that we have much more effectively, the famous stat that currently cars are 95% of the time idle, and so we eliminate that to some extent. So it’s a world where you see these machines moving around. They could be empty because they’re going to pick somebody up. They could be full with passengers who are facing each other. There could be a situation when some people are sleeping in that car as it takes them away. So it’s a different world, and every person has their own mental image of what that world would look like.

BRIAN KENNY: Bill, what’s your view?

BILL KERR: I think I could start with a personal side and then maybe a little bit along those lines as well. I think to your idea of opening up the newspaper, it would start at first being bliss, to free ourselves up. Then interestingly, the way humans behave, it would pretty soon become normal, and we would have some underlying level of grumpiness. You may have just cut my commute down to 20 minutes and I can read the newspaper, but that morning when it’s 22 minutes, I’m going to be like, “Come on.” So we will adjust quickly there. Then I think the challenge of us always predicting the future is that there’s so many things that are difficult to forecast. I always come back to electricity. One of the things that electricity allowed is for us to build really, really tall buildings. I don’t know that I would have at that moment of creation, I probably could have thought of the lights going down the streets and all that kind of stuff. I don’t know think I would have thought of Manhattan as being what that would look like in the future. So here you have a technology that could make non-routine transportation, so the idea that we can go from point A to point B, non-routine kind of work, make that super cheap, super effective, super easy. Then it’s like: what would that world look like? That’s one where we have to get in a time machine to get some sense of that.

BRIAN KENNY: Great. So Bill, I’ll stick with you for a moment. How does this relate to your research? I’m always curious about what prompts faculty to decide to write particular case.

BILL KERR: This is our very first case in an elective class called, Managing the Future of Work (MFW). WFW more broadly as well as in the course, thinks about how technology drivers as well as also demographics are reshaping the workplace and what companies are expected to do, how labor groups are going to be impacted, what’s the role of government and education, bringing all that in. I mean this is such a transformative technology. It’s the right here, right now that it brings in all of those elements.

BRIAN KENNY: Elie, what about you?

ELIE OFEK: For me, it’s all about the fact that this is transformational from the standpoint of the market and what are the implications from the standpoint of a breakthrough innovation, one that has so many far reaching implications and how the market gets to adopt this kind of a technology. So there’s the consumer side of it, the consumer psychology of thinking about innovation and whether or not there are barriers to adoption. Then at the same time this technology has not only consumer implications but customers are also businesses, and so what does that mean from the standpoint of the diffusion of this innovation? So I study many aspects of how technology and marketing come together, and so this is the dream case for a situation like that. There are so many elements that relate to thinking about diffusion and adoption and how marketing can get involved and what are the marketing implications for this especially when you take into account that the car industry is one of the biggest industries in terms of advertising and the way they try to impact people both on the functional level, emotional level and so on that this is a big disruption.

BRIAN KENNY: Bill, let me go back to you for a second. I alluded in the case that these technologies have been under development for a long time. Can you talk a little bit about some of the earlier innovations in this space?

BILL KERR: I think it’s helpful to remember or bring to mind that autonomous vehicles is about a system, a collection of technologies. It’s not about a single technological change. You can think of a progression of autonomous vehicles and what’s enabling this very futuristic vision we have versus what happens right now. Inside many warehouses and also in ports and other places, there’s a lot of autonomous vehicles. What we need to enable them to do is know where they are, sense things, and then know what to do. That setting of a warehouse or with a Kiva robot or something similar, almost like that earlier General Motors prescription, you can almost put down on the floor various things that says, “You are here,” and it can then run through this programming that says, “If I’m here, I need to get over there, and the coast is clear, then fine. I go that direction.”

Now let’s go a little bit further to what we envision for the future which is the three of us could decide after this we’re going to go to Upstate New York to go someplace skiing. What we would have then is an environment where we would be first navigating ourselves outside of Cambridge, highways. We would find more remote and rural areas along the way. We would come across stop signs that would be in different places. The stop signs might be half covered with snow. So you get into a much more complex sensing. It’s no longer about this thing that’s on the floor for you to register. You’ve got to be able to measure a lot about the world around you at all times. Then secondly is the programming logic, the “what I then do” becomes exceedingly complex. In fact, it becomes too complex for an if/then statement kind of structure to work. So that’s where we had to have advances in machine learning and artificial intelligence to help the computers just learn how humans would behave in a situation, appreciate the scenario that the sensors are providing you, and then act accordingly.

BRIAN KENNY: I thought it was interesting, too, both cases alluded to the fact that DARPA was one of the catalysts behind this whole thing. Can you talk a little bit about that?

BILL KERR: In 2001, there was a congressional mandate that they wanted one third of the military’s vehicles in war zones to be autonomous. Of course, there’s many points at which we made some mandate like that and then entrusted DARPA to help realize those outcomes. DARPA has under its credit things like GPS as well as also the internet and a bunch of relatively “small” innovations that have come out. You have those finger quotes there for listeners. So DARPA created a series of challenges that were in 2004, ’05, and ’06 that really helped push forward a number of the early technology advances that were in this space like LIDAR technology, also I think helped people to come together in those challenges and see what others were doing that was working and then adopt that.

BRIAN KENNY: Elie, your case talks about the five levels of autonomy in vehicles. I think we’re at level, what, two-ish at this point.

ELIE OFEK: I think building on some of the things that we’ve heard, there are five levels. In fact, there are six if you count level zero. The idea is that you start out with a level where the car needs a driver to do basically everything. Then you progress and say maybe there are a few things where the technology in car assists the driver, so you need the driver to make all the decisions and do all the driving, but maybe you get some warnings from the car that are telling you that you’re getting too close to the car ahead of you or maybe the car can do a sort of emergency breaking. That would be something like a level one. Then you progress over to a level two, and then you have scenarios where the car can adaptively figure out how to do the cruise control so you can be on the highway and then you can allow the car to decide at what speed to go in order to avoid hitting the car in front of you and adjust the speed and also to control… make sure that you’re in the right lane. So the car’s already taking over for you in those situations, but you still need the driver because the car itself can’t decide what to do and figure out exactly what to do. For example, if you’re getting into an intersection and things are getting messy there, then the car might not know what to do. That would be level two. Then you get to level three. Level three presumably, the car can do a lot of things on its own, basically take over, but the requirement is that the driver is still there since in a very complex situation the car will alert the driver to take over. So those are the first zero to three.

BRIAN KENNY: Level three is happening. Those are being tested in various cities. We’ve read about this, right?

ELIE OFEK: That’s right. That’s pretty much right. So first of all let’s say that level two is the Tesla autopilot that’s already on the market. That’s already a level two. Now, level three, there’s no commercially available vehicle yet, but those are being tested in various pilots. Now, most of the Holy Grail for autonomous vehicles would be levels four and five because if you start thinking about what’s the value proposition and if you start thinking about how do we think about not having a driver at all, because that’s where all the cost savings are going to come from and it relates to this future of work, then you want to be able to say I don’t even need the driver in the car at any point in time. So levels four and five are supposedly those levels where level four is under some circumstances you do not need the driver, and level five is all circumstances, all situations, no driver at all. Many of the car manufacturers had intended to get to that level and are working towards that level, but it’s proving to be a little more challenging than initially had been thought to reach that point where we can say we don’t need a driver under any condition or circumstance. But that’s what they’re working towards.

BRIAN KENNY: There’s a few different pathways to achieving this autonomy, right?

ELIE OFEK: The idea behind LiDAR is that the car is emitting basically light or photons. Those are hitting the objects in front of them and coming back, and then there are sensors that can sense that light coming back. Then they can build a 3D imagery of what’s in front of the car so that the car can decide how to navigate and drive. That’s been one of the technologies that has been promoted as one that would allow us to get to that level of four or five, and Waymo, as has been described, which is part of what was Google, is Google and Alphabet these days. But other players like Tesla are thinking that those systems might be limited especially under certain weather conditions. They are more taking the approach of cameras that are on top of the car or various locations on the car. These visuals that come in to those cameras, the shots that are taken, those are fed into the computer system of the car which can then analyze those imagery that are coming in through the cameras, and along with other technologies, like radar or ultrasonic vision as well, they will able to map out the terrain and understand exactly what’s going on. A third one is basically more on GPS and other technologies. So the players are differing in terms of what they believe will be the optimal level of information needed for the car to sense that environment as we’ve heard and then be able to make the right decisions. And they’re making bets. These are billion-dollar bets that they’re making on these technologies.

BRIAN KENNY: What is the ecosystem look like? Bill, your case goes into details about all the various players. I was amazed to see all the… It’s a huge ecosystem.

BILL KERR: It was a crazy mess at some level. I don’t think that’s so surprising if you think about this is a multi-trillion dollar market that could exist out there in the future, and many people want to have a place or a foothold in that space. If you get a list of the active players and this is a full enough space that many consulting companies or industry watchers will give every six months an update of where people see in this. It will start with Amazon and Apple at the beginning, and at the very end it’s going to be Uber or someone else in that space and all of the ones in between. Some that definitely get a lot of attention are the tech giants and Google and Waymo has been a big leader in this space. Let’s not by any means count out some of the traditional car manufacturers many of whom are trying to get into some space. General Motors in particular with Cruise and others have invested a lot in this space. There’s the ride sharing companies or the Ubers. They’re very active in this. Going to the cost portion of this, this is their biggest expense item in terms of providing localized ride service, so they’re trying to think about what that autonomous future looks like. Then you also have something on the order of 2,000 startup companies that are actively engaged in this space in some form or another and many partnerships that extend across companies.

BRIAN KENNY: Let’s dive in a little bit into the business implications of this because I think those are probably pretty significant. I’m wondering, Elie, in your opinion, which industry, I guess, or who in industry stands to gain the most or lose the most from a transition like this? It’s a huge shift in the economy if we were to make this full move.

ELIE OFEK: We’ve dived a little bit into the world of ride hailing services, so that’s one where you would imagine that these companies are thinking, the Uber, Lyft, that if they can eliminate the driver, that would reduce their cost structure. Interestingly enough, when you think about some of the other players and not coming from the ride hailing service but are coming at it from the standpoint of a Tesla or a GM, the car manufacturers, when they’re thinking, “What is the implication for us?” they’re thinking they might, themselves, launch their own ride hailing services when that comes about. So that’s one industry where you think could shift. Then if that shifts and the cost of taking an Uber or a Lyft, or if it’s another service that happens, drops to half or a quarter of what it is today, then you start thinking about, does it make sense to own a car as a consumer? Then the whole model of buying, leasing cars starts turning on its head, and it doesn’t make much sense to own… I can’t remember what the statistics is exactly, but a household today owns somewhere around 2.5, 2.6 cars per household. Maybe you only need one car because for everything else it’s so much cheaper to do ride hailing. So that’s one big industry that everybody’s thinking about which has to do with both the ride hailing yet the car ownership as well. The other one is logistics and delivery and shipments. If we think of trucking today, there are several million of these trucks around and hundreds of thousands of truckers. Can we reduce that? If shipping gets that much cheaper, that’s another industry that’s poised to potentially benefit. There, people are saying, “That makes sense. These trucks are driving on the highways. That’s much safer to begin with. So that may be less of a challenge than city driving in traffic, and so maybe we can perfect that and get that technology to work much sooner for trucks.” Yet to be seen. But that’s another big one that could shift. I’ll let Bill add on some more industries that could change.

BILL KERR: I think you’ve hit many of them right on the top of the head. I think one that I hear the most fear about is the traditional car manufacturers. In their future world, one BMW executive said, “We don’t want to become like Foxconn is to Apple. We don’t want to be just an equipment supplier into someone else who owns the software and the algorithms and that’s where the real profits ultimately lie.” But to this broader sense of things, the whole landscape could be changed in fairly dramatic ways. That being said, I guess the one thing I’ve yet to believe about this sector is that it’s a winner-take-all market. I think you could have multiple of these players come to a valued point in the future, a valued place in the future without having that only one platform is going to emerge as the dominant platform on there.

BRIAN KENNY: What does it means for jobs? Drawing on your research for the project you mentioned earlier, obviously we still need to make cars, so there will still be manufacturing happening, but you’re losing drivers. Tell me-

BILL KERR: This is actually how we spend most of our class conversations is on the worker themselves and trying to think through the implications that this has. A bit of background, the US’s largest job for a male is a truck driver. We have about three and half million truck drivers out there. To give reference, taxicab drivers and bus drivers combined are about one million so three or four times larger. In 28 states in America, it is the most prominent job. It’s overall a fairly well paid job. The median is around $40,000, $50,000, and a number of truckers earn more than $100,000. But this is where it starts to get a bit of the glass half full/half empty in that, while their salaries are like that, it’s a hard job to fill in part because people don’t want to be away from their families for such extended time and it’s arduous and those challenges. So we get into this question of will the technology aid the trucker in the future? Will it replace the trucker in the future? How long will this transition path occur? One of the starting debates in class is that if you look at, for example, Goldman Sachs’s forecast, they put the midpoint of the transition at 2043. I think everyone in this room is retired by 2043. That’s actually a little ways out in the future, and that’s the midpoint. Again, it’s already under negotiation. It’s a negotiated term between UPS and the Teamsters right now, the future of autonomous vehicles. So you get into this question of how are you going to manage that change in a way that helps that truck driver, and there’s many ways you could imagine helping the truck driver, if, for example, the autonomous part was the long haul segments and more local delivery is going to become the purview of the truck driver at either end. If you’re able to keep the trucking unions supportive of the action, well, you, as a company, are going to be able to move faster for autonomous vehicles because you have that support behind you. On the other hand, if they want to slow it down, they’re going to have a lot of power to slow it down. So we try to talk through those implications.

BRIAN KENNY: Does it create jobs anywhere? Are there new industries that spring out of this, or is it just redistributing where the work’s done?

BILL KERR: I guess we should acknowledge that there’s some estimates usually by industry groups but there are some estimates that would say even with the technology coming into place, you could have more truck drivers in the future. The thing is as you make that cost cheaper and cheaper, maybe we start shipping more things around the country period. A non-intuitive example is that in the introduction of ATMs into all of our lives, the numbers of bank tellers actually increased.

BRIAN KENNY: Wow.

BILL KERR: Now, it didn’t increase at the rate it was increasing before, but we didn’t see a decline. You saw the bank teller taking on more roles and responsibilities off around customer service and so forth, and because the cost structure of the bank got cheaper, we opened up more branches in other locations which led to the bank tellers becoming a bit more relevant. Let’s think first that we have to appreciate how all these logistic demands would change and how that could spill over. Then there’s so many jobs that we can’t even think about right now that would relate to this future autonomous world and either the entertainment that could happen in the cars or the retrofitting. There are so many actions that are going to be taken that we’ll probably be producing as many jobs in the segment.

BRIAN KENNY: Great. So it’s not like a gloomy picture on the job front. Elie, what about consumers? How are they feeling about getting into a car and not actually controlling it?

ELIE OFEK: Let me start by saying I don’t think all of them share your enthusiasm for getting in the car. When asked point blank, many consumers are somewhat fearful. I think one of the recent stats I was shown was that on a scale of one to 100 the confidence level in being in one of these was less than 50. It was something like around 36, which is pretty low for your confidence level in being in one of those. The other part of it is, what do consumers expect that government and regulation will do? So a lot of them are supportive of dedicated lanes for autonomous vehicles. A lot of them are supportive that there are certain areas like near schools and other public locations that they won’t be allowed without a driver in them. Even regulations that say there has to be the ability for an actual passenger to become the driver and take over. Now, of course, all of these things go counter to some extent to the big lofty desires that companies or businesses have in order to benefit completely from autonomous vehicles and then get the ROI from all of their investments in this.

BRIAN KENNY: Of course, both cases cite the fact that where the testing has been done, there have been a couple of accidents. There’s been some fatalities with some of these vehicles that have been tested. But it pales in comparison to the number of accidents that are caused by human drivers. In fact, I forget which one of your cases said this, but the accidents that had occurred in these tested vehicles most of the time it was a human that crashed into the autonomous vehicle not the other way around.

ELIE OFEK: Well, it depends on the eye of the beholder, I guess. I’ll just say one word, and I’ll let Bill continue. The point is that when you ask in class and you put it on a scale of one to 10, autonomous vehicles, do they have to be as safe or safer than current cars? Everybody believes they have to be safer. We’re holding them to a higher standard.

BILL KERR: We actually play a game that goes right along those lines because we start with… Most people the first thing out of their mouth, “They’ve got to be as safe as human drivers.” Then we quickly realize, no, they’ve got to be better. We hold technology to a higher standard. Then I start putting up on the board 99.999… How far do I have to go before you’re going to agree this is good? Usually, there’s like five nines past the decimal point. Everybody’s like, “Yeah, that’s good.” Then we discuss how there’s one billion car rides in America every day. So usually however far they’ve gotten on those nines, there’s still at least 1,000 accidents that are happening in America that day as a consequence of autonomous vehicles. So you start to unpack, then that has a lot of implications for insurance markets. Who are you going to sue? Do you sue the owner? Do you sue the software developer? Do you sue…? It’s a very challenging thing. Then there’s also a lot of, of course, delicious detail to think about in this transition from one to the next. You mentioned the case fact that many of these accidents are caused by humans, but they’re caused oftentimes by humans thinking the car’s going to behave the way a human behaves. But it doesn’t. It behaves the way the autonomous… So if you imagine one world in which we have fully transitioned to autonomous vehicles, there could be a level of safety and cooperation across those that is very challenged to have in a partial transition. Or another interesting thing is imagine that I am driving down the road and I really want to go three lanes over to this exit. I look over and I see an autonomous vehicle is on my right. I can just turn. I know that car’s going to stop. It’s going to let me go through just fine. So you could have this almost road rage or this differing behavior we could have towards the vehicles if we’re not careful.

BRIAN KENNY: It’s like bullying of autonomous vehicles on the highways.

BILL KERR: Exactly, exactly.

BRIAN KENNY: Speaking of the insurance companies, obviously there would have to be significant regulation of this industry which then brings the government into it which brings politics into it. What are some of the issues that you would see surfacing around the regulatory aspects of this?

ELIE OFEK: I think there’s multiple angles here. The first is, are we expecting the government to play a role in either facilitating or actually putting the breaks, no pun intended, on how fast these things are allowed to go? In that vein, it’s the matter of do we want them to give tax credits, tax breaks to those companies or consumers that adopt autonomous vehicles? The second part of this is, do we want them to do the flip…? So that would be a carrot. You’d give a carrot to those that adopt autonomous… But they could give the stick to those that stick on to non-autonomous vehicles and say, “You want to keep using your regular car, there’s an extra tax on you.” So that’s one angle. The other angle is, as I alluded to before, is the question of what are the regulations and rules on, do they have to go through dedicated lanes? Does there have to be a driver? How much testing is required before you get the sign off on your technology that it is safe to have as a level five? What is it? Is it hours of testing? It is years of testing under conditions that those regulators deem to be the right conditions that you tested those vehicles? I think the public is expecting regulations to do that. But up until now, yes, there has been thinking around this, but it’s not coordinated enough which could be something that’s deterring consumers from having that full confidence level in these. Then the third part of is, you were asking earlier about the technology to enable this, part of the vision around GPS-related ways that this could work has to do with the fact that we’d also have smart infrastructure, meaning that all the traffic lights and various elements along the road are connected and are able to signal to the car, “Hey, I’m a traffic light. I’m red right now. Hey, I’m a stop sign and so on.” The question is, who invests in that infrastructure? It could be that since the government, local government is the one that builds the highways, builds the roads, obviously there’s some federal assistance, is that going to be an investment that’s needed? Then where does the money for that come from? Taxpayer?

BRIAN KENNY: It’s getting complicated.

ELIE OFEK: It’s getting complicated.

BRIAN KENNY: What about the privacy implications? You’re mentioning GPS. People are already nervous about the fact that my phone is giving me away wherever I go, somebody’s tracking me. Would it only get worse in a situation like this?

ELIE OFEK: Yeah, absolutely. I think there are concerns. If you ask the person on the street, they will say that they feel like that the same type of tracking that’s happening to them online could be happening to them while in the car. That’s A. B, all of these things need to be connected. We have heard many scenarios and examples of hacking into computer systems. What if your car gets hacked? Could I then completely remotely decide what’s going to happen with your car? If I want to, god forbid, get somebody into an accident by remotely hacking into the car, is that going to be possible? We know any system, as much as you are going to try to make it safe, can be hacked. So that’s another concern.

BRIAN KENNY: This is for either of you. Are there other industries that you can think of that have gone through this kind of transformation?

BILL KERR: What we look at on the second day of the course is AT&T. It’s actually a two-part case with the first part being, I think, one of the closest parallel examples of this which is the Bell Telephone operators as we went from the human-centered switch into the mechanical switch. Again, to give some context, in terms of a percentage of the workforce, that was in percentage terms larger than truck drivers today. In fact, it was almost numerically-

BRIAN KENNY: Wow.

BILL KERR: … the same number of taxicab drivers we have today, and this was in a much smaller economy back then. That also went through a period of sustained transition from one technology to the other. It was being rolled out across regions. Our first case is during the Great Depression and they’re asking if AT&T is starting to slow down the rate at which it’s rolling out some of these technologies. We don’t know all that we wish that we could about this historical episodes. One of our colleagues at HBS, Dan Gross, is doing a lot of research around this. It does appear that there were transition costs. For a period of time this entry point for women into the workforce was not available. There wasn’t another channel that allowed for that. Of course, over time the technology was proving much superior, and AT&T as a whole was going to grow.

BILL KERR: What’s funny is that we catch them again at the iPhone’s introduction, so 2007 or ’08, and they realize that at that time that the data traffic that’s going over their network is too much. It’s growing too fast for them to continue to use the equipment-based approach that they had earlier. They have to go to software. So there’s this initiative that says we have to retrain 100,000 of our workers. We just love this. It’s the same company 100 years apart, and yet they’re needing to constantly make these adjustments in where they’re going.

ELIE OFEK: Yeah, right. Some people use the terminology of general purpose technologies to describe certain breakthrough technologies or innovations that have far reaching implications beyond just one specific industry or application. So examples would be the computer. Examples would be the internet where when the computer came out at first it was thought of as for a specific reason just like you would think an autonomous car would serve to just displace the way we currently drive, but it could have implications on industries that you never intended or thought of before and completely reshaped the way things are done in those industries.

BRIAN KENNY: I drive a Mazda, and Mazda’s branding is all about zoom-zoom. It’s about the feel of driving and how great that feels to get behind the wheel of that car. What does Mazda do in a world where I’m not actually driving those vehicles?

ELIE OFEK: That’s a big one. First of all if it moves to a world where there’s not even ownership, do you care about what’s taking you from point A to B? You could think of situations where just like today where you can decide on Uber if you want the more higher-end model to pick you up or not, maybe that’ll still matter to consumers. But you raise a very big point which is what have cars meant for us for decades? You take the ultimate driving machine which is supposed to be this feeling of power on the road and ability to signal some status-

BRIAN KENNY: That’s BMW right now.

ELIE OFEK: That’s BMW. Then you have “Love, it’s what’s makes a Subaru a Subaru.” So where do all those emotions go? It’s one where companies have used these kind of taglines and emotions and a desire to brand around certain imagery to sell their vehicles. It’s a big question about whether that goes away. Related to that is you ask people if they want to be in one of these cars and drive in one of these cars, but the question is, will that car drive the same way you do? Most likely they won’t as we’ve heard, and most likely they will be expected to be much more conservative, and most likely they won’t be cutting corners. So that experience itself of how the car drives when you are in control and you decided versus this AI and computing system is deciding could be somewhat frustrating. The scenario I play out in class is you need to get to your daughter’s play, you got only 10 minutes to get there, there’s traffic. Are you going to get frustrated if you’re in an autonomous vehicle and it’s driving below the speed limit, and it’s not going on yellows?

BRIAN KENNY: It’s like being behind a really slow person.

ELIE OFEK: Exactly.

BILL KERR: Already in Phoenix where they, of course, do a lot of this testing, there are humans that express this frustration about being behind the autonomous car that’s driving the speed limit, and they’re just driving the human driver insane behind it.

BRIAN KENNY: You both mentioned aspects of class discussions around this. I don’t want you to give away any big secrets, but any surprises that came out of conversations in class?

BILL KERR: I think for us a lot of the class is about the time that this transition’s going to take and appreciating that. One of the comparison points that has been interesting is, over the years, that transition time has been lengthened out. I think the first year I taught the case there was all like, oh my goodness, it’s not even 2043 for the midpoint. It’s going to be like 2020. We take a poll about this. I think last year probably the average transition time was another 10 years into the future. People are saying these edge cases that were described and the capacity to handle snow and all of these questions about people giving up their cars and the challenge, it’s going to be a lot more complex. Part of that goes to this general purpose technology idea which is that the car is so pervasive in our lives and we’ve so built our lives and our infrastructure and everything around the car and us being in that driver seat that the transition is so impactful if it happens, but it’s so hard to go from point A to point B given how pervasive this is in what we do.

ELIE OFEK: A few things that always are intriguing to see in class, one is people underestimate how difficult it is to actually drive a car and hence to create a car that can drive itself because you tend to think, what do you need to do? You need to steer a little bit, you need to break, and you need to press on the gas pedal. Right? That’s it. They don’t factor in all of these things that happen subconsciously when we’re driving. I think that once you put them in scenarios, they start realizing those. So that’s always an interesting moment to see in class.

BILL KERR: What’s intriguing about this scenario and what may be one of these ultimate questions is those edges cases and all that they matter so much, and there’s people’s lives and property and so forth that are at risk that it may be much harder and more complex because of the environment that this is working in.

BRIAN KENNY: So you’re both telling me I have to be patient. I won’t be able to read my newspaper while driving any time soon.

BILL KERR: You might have another zoom-zoom Mazda before this is all over.

ELIE OFEK: If you’re willing to do it in Phoenix, Arizona, on a test drive with one of these players, you can do it.

BRIAN KENNY: That’s my chance. Hey, thank you both for joining me.

ELIE OFEK: Pleasure.

BILL KERR: Thank you.

ELIE OFEK: I had a great time. Thank you.

BRIAN KENNY: If you enjoy Cold Call, you might like other podcasts on the HBR Presents network. Whether you’re looking for advice on navigating your career, you want the latest thinking in business and management, or you just want to hear what’s on the mind of Harvard Business School professors, the HBR Presents network has a podcast for you. Find them on Apple Podcasts or wherever you listen. I’m your host, Brain Kenny, and you’ve been listening to Cold Call, an official podcast of Harvard Business School on the HBR Presents network.