Adam Thierer is a senior research fellow with the Technology Policy Program at the Mercatus Center at George Mason University. He specializes in technology, media, internet, and free-speech policies, with a particular focus on online safety and digital privacy. He spoke with The Fuse’s Senior Reporter Matt Piotrowski about the liability issue surrounding autonomous vehicles.
Piotrowski: As you’ve noted and has been noted elsewhere, autonomous vehicles have enormous potential benefits for society, particularly with regards to safety. But one major question is who will be legally responsible when there is an accident. Could you explain your ideas on how to deal with excessive lawsuits so innovation and acceptance aren’t stymied?
Thierer: It’s important to understand that in the American legal system it’s usually the case that the party that has the most knowledge of potential harm is the one we usually affix liability to. Economists refer to this as the “least-cost avoider principle.” The person who is in the position to avoid harm at the least cost and who has the most knowledge of the subject is usually the one who is liable and is responsible for picking up the tab when things go wrong. Traditionally, that’s been the end-user, or the owner of the vehicle because not only do they own the vehicle, but they are in charge of driving it down a road, and if they don’t do that correctly, they’re going to be hit with liability when an accident occurs. Of course, that could all change and change quite rapidly in a world where autonomous systems and self-driving cars become more predominant. And if they do, it’s very likely that liability will gradually shift from the least-cost avoider being the fleet provider as opposed to the end-user. The fleet provider could be the manufacturer of the car, or it could be someone who has a fleet of autonomous vehicles and utilizes them to provide the public with service.
So those being held liable could be say the automaker or perhaps a company like Uber who buys a fleet of autonomous cars?
Exactly. It’s a well-known fact that Uber has made a lot of investment in driverless car technology. And clearly, they’re thinking about providing a fleet of robot cars to be at the public’s disposal at any given moment. In a world like that, it’s going to become increasingly likely we don’t own our cars. We’ll just lease them out by the minute. In that world, it would probably be crazy to think that we the passengers would have anything to do with liability in a crash. So, that’s generally speaking very, very good news. It will be a very disruptive trend that will certainly upend the insurance markets as we know them now. And it will have other profound ramifications.
In a world world of driverless cars, it would probably be crazy to think that we the passengers would have anything to do with liability in a crash.
What is of interest to me, and what I discussed in my recent Slate piece, is there is a question of how that liability shift will affect the actual diffusion of driverless car technology. Now, my primary hope is that there will be so few accidents in a world of driverless cars that this won’t be much of an issue. Hopefully we won’t even have to get to a point where we have to pre-empt common law and the torts and it just plays out in courts over a period of time, which could possibly be decades. We don’t know how long this transition will take. There will always be old codgers like me who will never want to let go of their steering wheel. The irony of my own position is that I spend a lot of time defending driverless car technology but I have absolutely no desire to have a driverless car. I love driving my car. It’s a passion of mine and has been my whole life, going back to my 1979 “Smokey and the Bandit” Trans Am.
I’m sure there will be a lot of other people who feel the same way you do.
The most interesting question about driverless cars is what’s going to happen the day we realize that they could bring us such an overwhelming public health victory that we have to make humans stop driving cars. One day regular cars will be such a novelty that we’ll look back and say, “My God, we actually let people drive tons of steel down the road at a high speed!” That will be a while down the road, but I think that debate will eventually come about; whether to allow humans to drive. Regarding liability, during this period of transition to driverless cars, if there is a situation when a couple of crashes happen that are particularly notable and concerning and the lawsuits start flying, there is a question of whether that could end up derailing what could potentially be an amazing technology. Again, the hope is that it won’t and the costs will be miniscule, even if there are a few lawsuits. But we live in a world where there’s a perverse incentive to file frivolous lawsuits at the first sign of any trouble. There’s no downside to doing so. There’s not really much of a cost to picking up the tab if you lose that lawsuit because we lack the so-called English rule, or the “loser pays rule.” I don’t know if that horror story will come to pass, but it has come to pass before, for something that was equally important—vaccines. I researched that experience and it was really quite troubling what happened.
During this period of transition to driverless cars, if there is a situation when a couple of crashes happen that are particularly notable and concerning and the lawsuits start flying, there is a question of whether that could end up derailing what could potentially be an amazing technology.
Beginning in the late 1970s and early 1980s, lawsuits started flying at the makers of childhood vaccines based upon what was some people’s assertions that were quite bogus—that these vaccines were causing problems in children or even deaths, when the culprit was in fact something else. But it didn’t matter—the manufacturers were getting hit by a ton of lawsuits. So they stopped making the vaccines. Congress and other policy makers at the time were hammering them to make more vaccines. But you can’t have your cake and eat it too. That problem became severe enough that finally policy makers went to the industry and asked what it would take to get the advancement of vaccines back on track so we could have these live-saving drugs for children to help prevent serious public health disasters. And the industry basically said something had to be done about the trial lawyers and the lawsuits. The solution Congress came up with was incredibly creative. It handles both arguments—it handles the problem of excessive lawsuits by creating a no-fault regime for the vaccine makers but it also handles the problem of the critics on the other side who worried about the handful of people who might be harmed by this.
Could you explain how the fund came about and why it was important?
So the fund that was created (it did have a small fee that was associated with it) was for the parties that were injured. They had to go to a special court—called vaccine court—and basically they had to make a claim against the fund, showing that their child had an injury or adverse reaction, or even worse yet died, because of a vaccine. But it was extraordinarily rare that this happened. But if it did, there was a way to rectify it without derailing all the life-saving innovation. I have to imagine that this is one of the most successful public policy and public health stories of recent times that is often not told. To the extent we go down that path, we could do something similar with autonomous vehicles.
What we don’t want to do is completely let the automakers off the hook and give them a get-out-jail-free card by completely immunizing them from any kind of liability.
What we don’t want to do is completely let the automakers off the hook and give them a get-out-jail-free card by completely immunizing them from any kind of liability. There are, though, examples of that in American law. People don’t realize that internet providers pretty much have a get-out-of-jail free card when it comes to online speech. Online providers are immunized from what any third parties post on sites such as YouTube, Yelp, and Ebay. The middleman can’t be held responsible by what any third-party says on their sites. There are some who criticize this law, but some like myself who praise it because without that immunity, a lot of those sites couldn’t exist. The internet, though, is mostly about speech, while vaccines and cars have a physical harm element. So the liability regime in this case has to be different.
This idea for the vaccine fund. Was it unprecedented at the time?
I’m still doing research on this for a longer paper, and I’m seeking out other types of regimes and I’ve found four, but they’ve all been subsequent to it as far as I can tell.
Has it been a model for other funds that have been developed?
I do not think so. But there is a similar model that is much more controversial, and that’s the one put in place in the early 2000s for firearms. There were a lot of lawsuits hitting firearms manufacturers for people engaged in shootings, and the lawsuits weren’t against shooters, but instead against the manufacturers of the firearms. The lawsuits suggested that the manufacturers should know that these weapons would create serious public harm. It was a big debate and Congress eventually passed a law that mostly exempted firearm manufacturers from liability. That’s been incredibly controversial, as you could imagine. I did not bring up that up in my Slate piece, but that law could be used as a model here. But I can’t find other examples like this before [the 1986 law]. That just could be because these tort problems we’re having with excessive lawsuits in the U.S. is a recent phenomenon.
How did you come across this vaccine fund? It’s not well-known.
I’ve been doing research on liability regimes for emerging technology for a while, and when I was doing this research, I was reading a nerdy textbook on law and economics and I found an example cited of this bill in 1986. I had no idea it existed, and I went and did more research on it. I was kind of surprised how little attention there’s been to this issue. I’ve also been looking for more research documenting its success or its problems, but I’ve been surprised that there hasn’t been much written about it. I can’t really explain that.
How long would this protection be needed for automobile manufacturers?
For the vaccine makers, once they made the drug and got it out, it was just a matter of whether they came out with another kind of drug. The difference for something like cars is that they could be on the road for a long, long time. It’s important to note, if a law like this is needed, it would have to specify what we are talking about liability for. There might be some types of truly egregious forms of product defects that are excluded. The firearms liability protection I mentioned does have a carve-out for something like that—if there’s a major defect in the product, they can still be sued. That might be the case here as well for autonomous cars. There would probably be a lighter touch so the most frivolous lawsuits couldn’t happen, but those types of details would need to be worked out. Is there some sort of a window or a timetable? Or would there need to be more sweeping, long-term immunity—say some 20 years down the road after a driverless car is on the road and something went wrong? Would it still be immunized? It probably would be. I don’t think that would change much over time.
Are you aware of any other ways to curb lawsuits against makers of autonomous vehicle manufacturers?
All of this assumes that the U.S. is nowhere near adopting a “loser pays” rule. It’s a non-starter politically for a number of reasons. The number one priority of trial lawyers is to stop any effort to change that default rule even though there has been a lot of evidence that the country would benefit from changing that rule. So my assumption is that you have to come up with second-best solutions if you live in a world where that perverse incentive to file lawsuits still exists. Beyond that, you could have other tweaks to the tort system, or you could have Congress change the legal standard of proof, or something like that. Again, my strong preference is to let the common law evolve naturally to respond to new circumstances and controversies. I just think there’s really one bad rule at the beginning of the process, and that’s the loser pays rule. That encourages a lot of frivolous activity. The loser is not accountable in any way. The problem is that a lot of the big companies in this space, like Google and Uber, could probably pay all liability costs. My concern is the Ubers and the Googles of the future that have not been invented yet.
The smaller, mid-sized companies may not want to be active in this space because of worry about liability.
There is a resource allocation decision that occurs in new fields like this. You have to decide right out of the gate whether you want to spend your budget on a team of engineers or a team of lawyers.
That’s exactly right. There is a resource allocation decision that occurs in new fields like this. You have to decide right out of the gate whether you want to spend your budget on a team of engineers or a team of lawyers. Google and Uber have already made a significant amount of investment in legal counsel, both internal and external. But small and mid-sized companies will not be able to afford this in the same way. They will probably be running on thin margins, so that is not a way to encourage life-enriching innovation from small and mid-sized players. That’s why immunity for internet providers has brought about such dynamism and small actors were not intimidated out of existence by the worry of lawsuits. I want the same thing to happen with driverless car technology and bottom-up innovation.
What do you think is the biggest obstacle for driverless cars? Are lawsuits the main issue, or are other factors more threatening?
I’d rank the threat of lawsuits very high. I don’t think it’s the primary impediment, however. One thing is the standards issues and the technical issues that need to be worked out. Those are not really policy issues, but standards issues, such as safety, infrastructure, and technical standards, can really hold back a lot of innovation. Those technical issues are broad and varied, but taken together could be impediment to widespread adoption. There are a lot of other infrastructure and technical barriers just within urban communities that could lead to a slower rollout in major cities. The sheer number of situations and obstacles that you’re confronted with when you drive through a city are very different than when you drive through the middle of the country. So, this is why I’m always arguing that the trucking sector will be the first one revolutionized by autonomous vehicles. Long-distance freight hauling will be disrupted by the new technology.
My hope is that the handful of resisters who hate driverless car technology won’t hold back the broader prospects for the change to occur.
You are talking about a massive revolution in the world of freight trucking. The federal government has authority over the interstate highway system so you won’t have this patchwork of regulations that you’ll have with cars in urban areas. There will be some communities that say “get those robot cars out of here,” or some simply won’t adapt to make driverless infrastructure and technology easier to use. There are also other cultural issues here. My hope is that the handful of resisters who hate driverless car technology won’t hold back the broader prospects for the change to occur. Hopefully, the autonomous systems are so intelligent that they can sense and avoid the dumb old humans like myself! Then you won’t have to worry about those who won’t let go of the steering wheel.