Explainer Episode 18 – Driverless Cars: Balancing Safety and Innovation
In this episode, Ian Adams lays out the current state of driverless car technology, discusses its implications for the disabled community, and breaks down the decisions federal and state policymakers are facing when it comes to regulating the technology.
Jack Derwin: Welcome to the Regulatory Transparency Project’s Explainer Podcase, which is part of RTP’s Fourth Branch podcast series. My name is Jack Derwin, and I’m Assistant Director of RTP.
Today, I’m happy to be joined by Ian Adams to discuss the technology and regulation of driverless cars. Ian is Executive Director of the International Center for Law and Economics and his work focuses on the disruptive impact of bridging technologies on law and regulation. In the past, he has testified in front of Congress about driverless car regulation. Thank you so much, Ian, for taking the time to talk with us today.
Ian Adams: It’s a pleasure to be here, Jack. Thanks so much for having me.
Jack Derwin: Absolutely. So to get us started, would you mind laying out for us the current state of driverless car technologies, as well as how it’s currently regulated?
Ian Adams: Absolutely. Well, so currently, driverless vehicles are of all kinds of different types. So there are different levels of automation in question. And they all represent different sorts of regulatory challenges. You don’t see them all over the roads, but there are some driverless vehicles that are working as delivery shuttles. There are others in Las Vegas that are, in very limited circumstances, moving people around. And then, of course, Waymo, part of Google, just announced that they will be doing services akin to a transportation network company like Uber or Lyft in the Phoenix area.
So you have these vehicles that are coming onto the road, to say nothing of the commercial side, where you’ve got trucks doing platooning, which is where they get very close to each other to realize efficiency gains associated with conjoined operation and not having to deal with wind resistance in each of the vehicles’ cases. So you see all sorts of different applications, currently. But of course, they’re not all over our roads. That’s still a few years off.
In terms of the regulation around these sorts of vehicles, you have different levels of governance. So at the federal level, you have the Department of Transportation, which has embraced a multimodal approach to regulating automated technologies, which is to say whether you are at the National Highway Traffic Safety Administration, which oversees passenger vehicles, or you’re at the Federal Motor Carrier Safety Administration, which oversees commercial trucking, they are all now taking a unified approach to interpretive guidance around automated systems. And that really is the relevant point when it comes to the federal regulation of these vehicles.
There is no explicit federal legislation pertaining to automated technologies. And therefore, existing safety standards are being interpreted in light of these technologies. And so at the federal level, each of those agencies is promulgating informal guidance to give direction to developers and to original equipment manufacturers.
At the state level, you have all kinds of different approaches. But those regulations are limited to testing requirements, licensing, registration, and liability. So most of the states now have something on the books about regulating automated technologies. But they don’t pertain to safety standards. Those remain at the federal level.
And then local governments are rolling out where they’re allowed to, because some states have actually preempted local action in this area. They’re rolling out controls around traffic, data collection, those sorts of things.
So you have these vehicles on the road. They’re not everywhere yet. But it’s going to be sooner than you know that we’re going to start to see all sorts of different applications of automated technologies on our road. And you’ve got three tiers of governance at the federal, state, and local level. And those are just starting to come into their own, as we speak.
Jack Derwin: Great. Well that’s super helpful. So this technology, and the regulation around it, tends to spark some pretty significant debates and those debates tend to pit public safety against innovation. So as it currently stands, one of the most significant shortcomings of existing law, when it comes to making self-driving vehicles available in a safe way.
Ian Adams: It really comes down to the lack of certainty that the current situation creates. The fact that you do not have federal legislation is going to lead to problems for, for instance, the judiciary as they’re asked to adjudicate controversies around automated vehicles. That’s not to say that you don’t have laws that can be applied regulations, that can be applied because of that interpretive guidance around their application.
But really, it’s an entirely new technology that will, at some point soon, require special treatment, special statutes. And so I would say that is a significant shortcoming. It’s not impacting us in a major way right now, when it comes to the deployment of these vehicles. But in the near future, it will. Because these vehicles are being developed using exceptions to existing safety standards to allow them onto the roads. And there are hard caps within statute about how many of those exceptions can be granted. And so we’ve got this situation where innovators are moving headlong to get these vehicles out onto the road to start to save lives. And they’re going to run into a hard cap unless Congress acts.
Jack Derwin: Thank you. So in all areas of regulation, laws and formal regulations tend to be accompanied by informal guidance, as well. If driverless car regulation is expanded, how would the prevalence of this guidance that’s quickly evolving and tends to be informal, concerning the development of driverless cars, impact their development?
Ian Adams: I would say that informal guidance, and I would add to this category executive orders, that the issue with both of them is that they’re subject to change really very quickly. And so while one administration, be it a governor or a presidential administration, may have one understanding of how they want to approach the technologies, of course that can all be changed overnight when a new administration comes in.
And I say this because, of course, we’ve got an election before too long, and these are very real considerations. And when it comes to developing these new technologies and making bets about investment, the lifecycle, or rather the product development cycle, happens in a — on a time horizon of five to seven years. And so that does not align particularly well with the changes that we see in our political classes. Because once the politics changes, that regulation can change so quickly, and guidance can change so quickly, can cause real problems. And it chills investment.
On the upside, though, informal guidance, of course, when it is achieving outcomes that you find preferable, where in this case it happens to, from my perspective, because it’s been promoting the development and deployment of automated technologies, can be really handy. Because these vehicles have been essentially by fiat, fit into existing regulations. And therefore, as the technology evolves, the guidance can evolve with it. And that’s really very helpful.
But you can see, Jack, of course that there’s a tension between predictability and the ability to actually get these sorts of vehicles on the road. Because if you’re just waiting for Congress to act, or you’re waiting for an agency to promulgate regulations, you’re going to be waiting for years, and years, and years. And that’s not very helpful when it comes to getting burgeoning technologies out to the public.
Jack Derwin: Right. That’s really interesting. And so, I’ll turn now away from maybe the typical things we think of when it comes to regulation. And as such a new and still constantly evolving technology, do you think our institutions, such as law enforcement and our judiciary, are ready to interact with the new technology?
Ian Adams: Well, of course, there are big questions about liability when it comes to automated technologies. We’ve got questions about whether our existing system, that is based around negligence, is going to be a good fit for what might end up being a lot of product liability sort of situations.
So I think that our courts, because they have general jurisdiction in many area — in virtually all areas, of course, they are prepared insofar as they’re prepared to handle any sort of controversy. But they’re certainly not prepared when it comes to the background education on the nature of the technology. And that probably is just a matter of exposure.
So over time, these institutions will be better equipped to interact with the technologies. And until they are, you’re going to start to see some really aberrant conclusions from courts. You’re going see inconsistent enforcement from law enforcement agencies. And while regrettable, I think that’s probably inevitable. Just as you integrate new technologies anytime in a major way into a society, you’re going to see inconsistencies and you’re going to see flaws in the way that technology is integrated.
Jack Derwin: Right. Thank you. So a lot of the focus on this technology tends to be on the commercial aspects. But one often overlooked area is the benefits they might have for folks who aren’t able to drive, such as people who are blind or otherwise disabled. Can you talk a little bit about how this technology might impact the lives of people like that?
Ian Adams: Well, I think that’s probably one of the areas that gets the least attention that should get the most attention. Because you have underrepresented parts of society, under-involved parts of society who, by virtue of the fact that they have some sort of a disability, simply are not able, yes virtue of technological limitation, they’re not able to do the things that so many of us are able to take for granted.
And I think that’s one of the great things about automated vehicles. That if you are a blind person, you can now conceivably get a license to operate an automated vehicle in a safe fashion that will allow you to travel without relying on others. That is going to be fantastic for huge groups of people. And it’s going redound to the benefit of all of us.
So not only are the commercial and safety benefits fantastic, but I guess I am really looking forward to the ways of which the reintegration of whole parts of society is going to improve our lives. And it really is just a technological question that we’ve been unable to answer that’s kept them on the sidelines.
Jack Derwin: That’s really, really interesting. Thank you. So sort of on a similarly overlooked note, a society filled with driverless cars all over the roads would necessarily mean that the companies operating them would have access to an incredible amount of data about the occupants of those cars. Do you think that reality would change how consumers view data collection? And would that reality necessitate any new regulations focused specifically on the data elements?
Ian Adams: Well, I tend to think that consumer acceptance of a technology is ultimately the greatest barrier to its success or failure. So when we think about automated vehicles, we’re often thinking about the success or failure of the technology on the basis of its technical evolution. Which, I mean, necessarily you have to have a vehicle that works before it can be experienced by anyone. But ultimately no. The success of failure of automated vehicles will come down to the extent to which people are willing to interact with them, because that is the necessary component for any firm to invest in a major way in a technology.
When it comes to privacy, I think it’s the case that standards will evolve, as we saw with the introduction of the camera, where there were widespread concerns about public likeness, or someone’s likeness being appropriated in public because of the camera’s development. Well, of course, society has evolved its standards around likeness and the capturing of images in public, and expectations of privacy have evolved along with it.
Well, with automated vehicles, all of the sensors that are on these vehicles that are collecting all sorts of data, both internally based on the passenger or operator of the vehicle, but also externally based on what the sensors are picking up around them, I think it’s the case that as you see the application of that data in different ways, different concerns that people have will come to the fore, and that could inform policy.
I think the big mistake would be to try to anticipate the nature of those concerns and prescribe particular behaviors or sensor types ahead of time. Because that’s really trying to read into a crystal ball in a way that cannot be done effectively without seriously circumscribing the natural evolution of the technology in a way that could lead to the highest and best societal use.
So do I think that there will probably standards around data collection and data use when it comes to automated technologies? Yeah, probably eventually. I think that now is not the time to attempt to guess about what the public really wants from those standards or regulations.
Jack Derwin: Great, thank you. So without asking you, as you said, you peer into your own crystal ball, looking forward though, do you have any sort of prediction or sense of when this technology really will take over and similarly, when a piece of federal legislation might come into place?
Ian Adams: Yeah. Well, two good questions that, if I had the answer to, I would be a very, very wealthy individual. But to your first, I think back to five years ago where auto makers had what had proven to be really optimistic claims about the deployment timeline and the development timeline associated with this technology. And there’s, I would argue, been an overcorrection. And so we’re sort of in a trough of disillusionment where some are saying, just this technology is not realizable. It is not going to be on our roads. It is not going to change our lives in the way that was discussed five years ago.
I would say that the misunderstanding has really been around the types of technology. Because we are going to see deployment. It is not going to be the level five vehicles that can do everything in every circumstance. These are going to be applications of automated technologies in limited communities with very limited design domains, which is to say routes.
You are going to start to see those things in the next, I don’t know, year I would say in a meaningful way. Because you’re already seeing trials in large communities with the general public. When will it hit everyone? When will they become ubiquitous? That’s a hard question to answer.
I will say that one of the stumbling blocks that will prevent this technology from becoming ubiquitous really is to your second question, one about when Congress sees fit to legislate in this space. And to develop specific law related to automated technologies. And while we have seen a number of bills move forward in Congress, and there is one under act of consideration right now, I think the political winds are in motion and that Congress isn’t going to act anytime soon. And so I have to hope that they’re going to get it in the next couple of years and that we will see something move forward. But in the near term, there really isn’t much hope for action from Congress on automated vehicles.
Jack Derwin: All right. Well thank you so much, Ian, for taking the time today. I think it was really interesting to dive into a very much talked about new technology but not one that necessarily everyone always gets into the weeds about. So we really appreciate it.
Ian Adams: Yeah. Happy to join you, Jack.
Jack Derwin: Thank you for tuning into this episode of RTP’s Explainer Podcast. Please check out our website, regproject.org to learn more about this issue, as well as a host of other regulatory topics.