Conversation #1: Harry Lightsey, Executive Director of Emerging Technologies Policy at General Motors (GM)
See all Mobility Futures conversations here.
As part of the “Big Three” in U.S. automakers, General Motors (GM) has been aggressively working towards putting fleets of self-driving cars on the roads in 2019 (learn more), from major acquisitions to filing exemptions for road testing new models.
I met with Harry Lightsey, executive director of Emerging Technologies Policy at General Motors to discuss GM’s vision of the future of auto use. The discussion has been edited for length and clarity.
June 2018, Washington DC
Kathy Zhang: Harry, so you have a big title. What’s under the emerging technology portfolio, and more specifically, the parts that you think could hit the road in the next couple of years?
Harry Lightsey: So generally speaking, emerging technologies policy, it’s kind of continuously being redefined as new technologies start to emerge. We cover a kind of urban mobility — the various solutions that the GM business is looking to provide to address issues around urban mobility in the future, and that includes self driving vehicles. We also do a lot of work around connected vehicles, and that goes anywhere from talking to the FCC about spectrum needs to dealing with data privacy and cybersecurity and those types of issues that are part of the development of those vehicles.
KZ: Something I hear a lot from policy makers at all levels of government, is that, this is all happening so fast — like drinking from a fire hose. Are we seeing a Transportation 2.0 of sorts, with a convergence of different types of technologies?
HL: Two things on that. First of all, I think the automobile industry transportation, if you look at it from that prism has been kind of defined for the hundred years: that it’s existed by a vehicle powered by a gasoline combustion engine, driven by a person, and owned by person. And all three of those are changing at the same time. Now we’re seeing commercially buyable powertrains battery, electric vehicles, hydrogen fuel cell vehicles, there are even the others beyond that, but those are the main ones that we’re seeing right now.
And we’re seeing self driving vehicles, so we’re taking the human out from behind the steering wheel, and we’re changing the ownership as well because now it’s not about a person owning a vehicle. It’s about a ride sharing and car sharing and pooling and all of those different combinations that don’t look like individual ownership anymore.
I think it’s kind of the fact that all three of those are happening at the same time. It tends to overwhelm people, but on the other hand,
I think the media has kind of overplayed the rate of change. I don’t think it is changing as fast as maybe some people feel like it is.
I think you know, we’re still at the very, very beginning of actually deploying these technologies, and I think what we’re going to see, you know, in the next five years, we’re not going to all wake up one morning and all the cars out there and driving around by themselves. I think what we’re going to see is deployments in urban cores where you have the population density to generate a sufficient demand for ride sharing services. Also where vehicles can operate in a geofenced and defined area because that’s where the technology is.
I think that’s what we’re going to see for the foreseeable future. So it’s not going to be everywhere, and it’s not going to be all the time. The public will have the opportunity to start to experience the technology, and the technology will start to develop. And getting the trust of the public for driverless cars as a reliable and safe means of getting around. I mean, those are the two overall goals that we can try to achieve in the next five years.
KZ: I’m curious about this model of private ownership versus fleets versus ride sharing. The projections on deployment timeline and fleet models are all over the place. What does GM think about different ownership models, and what sort of time horizons is GM trying to model out?
HL: For GM’s purposes, I think what we think is that in certain parts of United States and in the world, individual ownership is still going to be the right solution for a long time to come, maybe a generation or two to come. Because of congestion, emissions concerns and other things, people are looking for different kinds of solutions. In a dense urban city, it’s a big time hassle to own a car: you have to pay taxes on that, you have to pay to park at, you have to pay insurance on it, and you have to maintain it. If you could solve all that and still be able to get around when and where you want to, it’s a great equation to have.
The key to that is you have to be able to get where you want to when you want to get there, so you have to have that independence and that freedom of mobility. As long as people are within an urban core, and all they’re doing is wanting to get around within that urban core, we can meet that kind of demand. But if they want to go take a trip to the country, they’re going to want a access to an individual car.
KZ: So you talked about connectivity. There might parallels in AV coverage being like cell coverage, and over the years you see a graduate expansion in coverage in the U.S.
HL: I don’t think any of us know the answer really. We do know that traditionally the car parked turns over very slowly, and the average ownership of a vehicle in the United States is over 11 years now. And that’s the average. You can see how it’s going to take a while for any kind of new ownership paradigm to start to hit the fleet in appreciable numbers.
On the other hand, disruptive technologies tend show much quicker rates of adoption than expected.
I worked in the communications industry, and the adoption of the Internet and the adoption of wireless phones was much quicker than expected by the industry, based on traditional model and traditional expectations.
KZ: What was wrong in the model?
HL: One thing was the cost of the technology. My company in telecomm, we had done all those market surveys and research, and we had determined that what people really wanted was to have a wireless phone that they could use for voice conversations, and that they’re willing to pay $80 for it. So we were busily working on trying to figure out how we were gonna have a cell phone that cost 80 bucks, and Steve Jobs and Apple came up with a cell phone that costs a thousand bucks. And then they figured out a way to provide it to customers, heavily subsidized by the network provider. You know, it flipped the whole paradigm. So that’s the example that I would use.
Whether something like that can happen in the automobile industry, where you’re talking about much more significant amounts of money and all of those kinds of things — it’s questionable. There are interesting articles out there on everything from ride sharing, access to transportation to a fractional ownership of vehicles, and all of those different types of paradigms. The vehicle is, for most families, probably the second most expensive purchase they’ll ever make, with their home being maybe the number one.
You’re paying a significant amount of your savings or income for a vehicle that, for most of us, sits in a garage or a parking lot at work or a garage at home 95 percent of the time.
KZ: I don’t think people fully understand all the features and capabilities of their current model cars, let alone future autonomous capabilities. I’m curious as to how GM is thinking about the public education effort to build consumer confidence, trust, and acceptance. What might that process look like?
HL: One of the things that we’re working very, very hard on is to kind of emulate the telecom industry, if you will, and make the [vehicle] user interfaces as intuitive as they have done. There’s a lot of effort and research being conducted in the automobile industry into how can we present this technology in a way that’s intuitive and easy to figure out how to work.
For example, we have this system called Super Cruise. It’s kind of like Tesla autopilot, if you’re familiar with that. With Super Cruise, if you’re on a divided lane, limited access highway, you can turn it on and you can take your hands off the steering wheel and take your feet off the pedals, and the car will steer in the lane that you’re in for as long as you want to have it engaged. If comes up on a slower moving car in front of it, it’ll slow down.
If the car in front moves or gets out of the way, it speeds back up to whatever speed you set it at. There’s some very sophisticated technology that enables to do all that, but the user interface, how you turned it on and how it shows you that it’s functioning and when it shows, when it lets you know that it’s time for you to take back over control of the vehicle — all of that is very intuitive, a matter of three different colors of blinking lights on your steering wheel. So, you know, I think that’s the future of how to make it so that people can take advantage of the different features in their car.
KZ: Safety is the major value proposition of autonomous vehicles, but two key questions seem unresolved to date: first, the consensus on how safe AVs should be, and the question which precedes it, how do we know how safe they are?
HL: We know that people are going to get behind that wheel when they shouldn’t, when they’ve had too much to drink or something. We know that people are going to attempt to drive when they shouldn’t, when they’re looking at their cell phones or they’re distracted. And we’ve seen the toll all that takes.
The expectation is not going to get any better — humans are doing what they’re capable of doing. There’s not going to be anything that suddenly makes humans capable of driving, you know, 20 times safer than they already are. On the other hand, if you look at where we are with AVs [safety], it’s a curve and it’s going up.
We’re not at a level of the human being driver yet. We hope we’re approaching that very quickly.
But the key thing about that is once it passes that line of being as good as a human being driver, it continues to improve and the improvement line is kind of infinite. There is no end to it, and it just gets safer and safer and safer. That’s the potential.
So the question is, when is it that these vehicles are safe enough to be on the road by themselves? We would say that at least when they’re is as safe as a human driver, because at that point you’re not losing anything in terms of safety, but you have the tremendous potential to get better as you go forward. Once we can establish that we’re as safe as a human being driver, then we should be allowed to put these vehicles on the road and to continue to develop and improve the technology. So that’s kind of point A.
So point B is, how do you establish that [AVs] are as safe as a human being driver? That is an exercise that we’re all engaged in. Ultimately, we’ll figure out with regulators and others — it’s really NHTSA’s charge to determine when that is. Two things things have to happen so the NHTSA promulgates these federal motor vehicle safety standards, the FMVSS. As they’re formulated now, they anticipate that there’s a human driver behind the wheel. So we could not put a vehicle that didn’t conform to those [standards] on the road unless we get a approval from NHTSA to do that, and they’ve made it very clear that they’re not going issue that approval until we establish that the system that we are putting in to replace whatever FMVSS we can’t meet, is as safe as a vehicle with that system in place.
So we have to prove equivalent safety. We have filed a petition for relief for a vehicle that we would like to deploy sometime next year. A picture of it is on our safety report — it doesn’t have a steering wheel pedals, so obviously it can’t comply with the current FMVSS. We’ve filed a petition with NHTSA basically saying, here’s why we think this vehicle is as safe as a vehicle with a steering wheel and as safe as a vehicle with brake pedals and all of that. And we expect to go through that process with NHTSA.
KZ: Is GM going to stand behind full liability in case of failure?
HL: We think that until we establish otherwise, we believe that the current law is adequate to deal with it and that we shouldn’t try to anticipate what the legal issues may be until we actually experienced those and we know what they are. We certainly intend to be responsible for the vehicles in the systems that we put on the road just as we are today, and we’ll stand behind that.
KZ: What sorts of collaborative efforts are there in the auto industry or with tech newcomer around cross-cutting issues such as data sharing, safety, and cybersecurity?
HL: What we’ve seen today is kind of indicative of an industry that’s in its infancy. You have a lot of new players and some established old players that are working very hard to try to develop these technologies as rapidly as they can, and I don’t think there has been a lot of collaboration or working together, to be honest with you.
I think that as the industry continues to mature, you’ll see more of that, as people begin to put products out on the road and it’s no longer such a huge secret about what their product is capable of doing or not doing, you’ll start to see more people looking at more common platforms and stuff. There are efforts underway in kind of the typical standard setting bodies like Society of Automotive Engineers (SAE), to start to come up with standards, but those are just very early on stages of being developed. I certainly hope to see a lot more of that as we go forward, but I don’t think to date anybody would say that there has been much of that.
KZ: Do you see the aviation model — in the way that different airlines come together and share data on failures, crashes, edge cases — as something that would emerge from such collaborations going forward?
HL: It’s possible — I would say that the airline industry and the airplane manufacturing industry is much less competitive than the automobile industry. You’re talking about globally, three major manufacturers and they’re producing a new model once in a generation. Automobiles are being produced by hundreds of companies globally, and they’re iterating a lot faster than that. I think [collaboration in the auto industry] is magnitudes of degrees more difficult than the airline industry model. That’s not to say it couldn’t happen or some version or flavor of that couldn’t happen, but I think it would require a whole lot more work.
KZ: You brought up earlier vehicle-to-vehicle (V2V) communication and vehicle-to-infrastructure (V2I) communication, which would create clear safety benefits. What do you see as the role of the private sector, whether it’s the auto companies or telecomms, to try to accelerate this technology?
HL: So it’s interesting — the early concept of a B2B communications was come up with in the 1990s and then in the early 2000s. Folks convinced the FCC to allocate abandoned spectrum to the development of the technology and a lot of work has been going on to develop the technology. Probably about four or five years ago, the industry pretty well established through field trials and other things that the technology worked and was capable of having a tremendous impact on safety as you indicated.
Unfortunately, there hasn’t been a kind of a large impetus to deploy the technology. It’s one of these kind of chicken and egg situations where all the cars have to talk to other, regardless of whether they’re a GM car, a Ford car, Toyota car, they have to talk to each other and they have to be able to understand the language that each other is talking so that the safety benefits can be created. Unless everybody deploys the technology, you know, GM doesn’t get any competitive advantage by deploying the technology by itself — all of the car companies have to deploy.
For those reasons, GM supported the proposed mandate that NHTSA came up with to require that all the companies put it on their cars. Now that hasn’t happened, and that rulemaking is still ongoing at NHTSA, so we’ll see where that goes. In the meantime, GM made a good faith commitment and actually deployed the technology on its Cadillac CTS beginning in 2017 and then a couple of weeks ago, GM announced that it was going to expand that deployment to another kind of like vehicle in 2023 and then ultimately to the entire Cadillac portfolio.
Those are all voluntary commitments without any assurance from any other automaker that they’re going to do anything. Toyota made that same kind of commitment, that they would begin a deploying the technology in their fleet in 2021. So you’v got GM and Toyota out there now, who have stepped out voluntarily, if you will. I think that’s a very strong, good faith commitment on those two companies part to begin the momentum to realize the technology and hopefully other car makers and OEMs will step up and do the same.
I don’t know how closely you follow all of this, but now all of this work that I’ve been talking about has been based on a communications protocol, basically a wifi-based protocol that was actually created in the 1990s when the technology began to develop. Most recently, the cellular communication standard setting body, the 3GPP, certified the standard for point to point communications.
In the latest release of 4G LTE, there are a number of automakers and communications companies that have said that they believe that this cellular-based standard offers a much better path forward than the DSRC wifi-based standard. Right now what we do know is that the wifi-based DSRC does work. There is going to be testing done on the cellular V2X technology beginning very soon, but chip sets aren’t even available yet for that technology. So, you know, that testing has to occur and that has established that that technology can work because people’s lives literally are going to depend on it.
And before we know whether that technology is viable or not, what we’re saying at GM is, we can’t afford to wait for all of that work to be accomplished when we have an option or a solution that works today that we know about. That’s why we are going ahead with DSRC, and we’re not waiting for the cellular V2X to develop. If the cellular V2X proves out to be a more elegant solution, a better solution in there, we’ll figure out a way to migrate to that. We don’t think that we should wait for that to happen when people’s lives are at stake, and we have a solution that can be effective today.
