Podcast Transcript: The Legal Impact of Driverless Car Technology with Akerman’s Gail Gottehrer

Topics: Business Development & Marketing Blog Posts, Efficiency, Law Firms, Legal Innovation, Podcasts, Thomson Reuters

podcast

Hello, and welcome to the Legal Executive Institute podcast. Today, we begin a series of two podcasts around the legal impact of driverless car technology and how law firms can meet this coming need. Speaking with us is Joe Raczynski, a legal technologist and futurist for Thomson Reuters Legal. Joe will be speaking to Gail Gottehrer, a partner at the Akerman law firm. In Part 1 of our podcast, Joe and Gail will discuss driverless technology, where this technology is, and who are the major players in this field. In Part 2, they’ll discuss the benefits for law firms and how law firms can seek opportunity in this area.

Joe Raczynski:          Thanks Gregg. Gail, welcome, welcome. We really appreciate you joining today for this discussion.

Gail Gottehrer:         Thanks Joe, it’s a pleasure.

Joe Raczynski:          Great. Well, I am very excited to chat with you about this fascinating topic, no question. The topic of the day is autonomous vehicles, self-driving autos, or driverless cars. Recently, I actually saw an article in Forbes that was published. It said that by 2020, this is, I guess, about 2½ years away from now, there will be roughly 20 million self-driving cars on the road. Do you believe that?

Gail Gottehrer:         I think it’s a nice idea, but I think it’s a little optimistic. I think the technology is there. That’s not the impediment. The impediment is integrating these vehicles and the technology that we’re able to design into our existing world, so our existing completely not prepared for autonomy roads, toll booths, bridges, things like that, insurance schemes, laws. The impediment is not the technology, but it’s people, process, laws, the whole scheme that these vehicles have to fit into.

Joe Raczynski:          That makes sense, totally. Maybe taking a step back and helping define this space. I remember as a kid going on a family beach vacation, dad driving the family truckster, mom took the helm as the navigator, and the kids were arguing in back. It was good times. To help ease the pain of that elongated trip that my dad and the family were taking, he tapped into this new cutting-edge technology that we used to call cruise control way back when. Gail, in helping us understand this autonomous technology, does cruise control count, and further, are there industry classifications or standards around automation, I guess?

Gail Gottehrer:         Sure. Cruise control you can think of as in the evolution of this kind of technology. Cruise control was a very early step and what we’re talking about with autonomy is really further down the road than that, but that was kind of one of the building blocks or the early precursors to things now that are in cars that people probably take for granted, or don’t view as autonomous components would be things like driver assist, or if you have a car and if you’re veering out of your lane something beeps, or something vibrates in the steering wheel to alert you to it, or the lights flash, something like that. That would be kind of the modern day equivalent of kind of what cruise control technology was there. There is a range and there are different ways of looking at it, but the most accepted classification I think and the easiest for people to understand is from SAE, which is the Society of Automobile Engineers.

For that, we go from Level 0 to Level 5. Level 0 is no automation at all, so even less advanced than what your dad was driving on that happy family vacation. Level 1, the driver performs all the driving tasks, so that’s what people think of generally as driving. Level 2, you have one or more of the driver assistance systems are at work, so they’re helping out, so like cruise control, lane assist, technology that will keep you in your lane if you start to veer, things like that but the driver is doing the dynamic driving tasks. You have to have your foot on the pedal, your hands on the wheel. There’s software, they’re algorithms, there’s technology in the car to help you, but you’re still driving. Still what we kind of think of today as traditional driving. When we get to Levels 3 to 5, this is where we really have automated driving systems monitoring the driving environment and this is more of what we’re talking about when we start talking about these kind of autonomous vehicles and I think what you were referring to in the Forbes article.

Akerman’s Gail Gottehrer

Level 3 you can kind of think of as conditional automation, so there the automated driving system controls either some or all of the systems of the car with the expectation that the driver is going to respond if there’s a need to intervene. At a certain point, it could be cruise control or beyond, where you could take your hands off the wheel, your foot off the pedal, but the car would alert you if, for example, you got to a place where the weather was bad maybe and the sensors on the car got fogged up, or there was ice, something that the car then knew it wasn’t able to distinct or interpret properly and needed your help, and needed you to intervene. Something would alert you that you need to start traditionally driving again, manually driving. The issue there, though, and this is why it’s complicated at this Level 3 from an insurance perspective and a legal perspective, are questions of who has the responsibility to be in control at what point and if you have an accident, whose fault is it?

Were you driving? If you didn’t take over in time when that warning came on telling you to intervene, did you act reasonably? Some estimates say it’ll be four seconds that you’ll have to take over and a four-second response time would be a lot easier for someone in their twenties than somebody in their eighties. How we deal with that is going to be another issue going forward in Level 3. Level 3, for that reason again, is really why it’s viewed as the most complicated level. Then we move onto Level 4 and there are some companies that are actually thinking about skipping Level 3 altogether because of those complications and going directly to 4. Level 4 is kind of thought of as high automation and at that level the automated driving system controls all the aspects of the dynamic driving tasks and the car can drive even if the person doesn’t respond to a request to intervene. The car may tell you, “Take over. Get back involved in driving,” but if you don’t, the car will be able to get itself through the situation.

Likely, it will slow itself down, move you to the side of the road, stop itself, but it’ll continue even if you can’t intervene and you don’t respond to its request. Then finally, where we get to what we all dream of is Level 5. Full automation, which is all the … The automated driving system does all the driving tasks that a human could do and under all road conditions and all environmental conditions. At that stage, the person in the vehicle is essentially cargo. You’re not driving, you’re not expected to drive. You’re just a passenger and you’re just an occupant of the vehicle, you don’t have to operate it at all. That’s where you may have seen some of these pictures online of people potentially sleeping in their cars and these vehicles becoming more like a mobile living room with infotainment, and leather kind of couches, things like that. That would be the vision of Level 5.

Joe Raczynski:          Wow. I actually really like that vision, so thank you for going through those different five different levels of the SAE.

Gail Gottehrer:         Sure.

Joe Raczynski:          The next thing that you sort of talked about briefly was the technology behind it. Over the last few years, I’ve personally seen pictures of cars with huge, bulbous-like spinning eyes on top of the car, and sensors that simulate sort of pop out of the sides of the car. At a high level, can you touch on how this technology works?

Gail Gottehrer:         Sure. Essentially what it is, is while they’re not necessarily the most aesthetically pleasing, that eye spinning on the top of the car, these vehicles operate by having a tremendous number of sensors around them so that they can sense what’s going on in other lanes, with environmental conditions, so it’s kind of… think of it as a 360-degree view being taken of your surroundings and that in combination with all the technology in the car, the sensors, the lidar, which is sort of like radar, think of that way, it’s constantly getting information about the car’s positioning where you are, how fast you’re moving, how fast other cars are moving, where the lane markers are, if there are obstacles in the road.

It can sense where the traffic lights are, if there are stop signs, and there’s a lot of technology out there. IBM has some excellent products where they can read the roads and they can tell, for example, that something is a stop sign by seeing its shape and the number of feet it is from the curb based on local ordinances of where stop signs are placed. All these sensors and all this technology and that spinning eye is taking in your whole environment, processing it, and giving the information to the car to operate.

Joe Raczynski:          Fascinating. Does it use a ton of memory, I’m assuming?

Gail Gottehrer:         Yes, it uses a tremendous amount of memory and the amount of data it collects is also staggering. There have been some estimates of like four terabytes of data a day and even without automation, there’s also V2V and V2X, so vehicle-to-vehicle communication, vehicle-to-infrastructure communication. There’s a tremendous amount of information going to your vehicle under these systems, from your vehicle either to part of the design of smart cities, or to other vehicles. The vision being that, eventually if all cars are either at some level of automation or at least connected, that every car will sense each other and you’ll have fewer if any accidents because cars will all sense each other and be able to react to each other and avoiding hitting each other.

The problem is, that these vehicles are going to be expensive for quite a while and people, the statistics say, own their car at least 11 years on average. It’s going to take a while before people can be in a position to have an autonomous vehicle, so what you’re going to have are roads that have a mix of older cars. You’re going to have things ranging in the Level 1, potentially Level 2, Level 3, Level 4. You could have all those on the road at the same time and that’s where it gets complicated, because your car may be able to take in all this information, communicate, but other cars won’t be able to communicate with it.

Joe Raczynski:          Wow. No, that is amazing. As we start thinking about the players in the US that are sort of tearing up the tracks and excelling forward here. Gail, can you briefly go through the top companies and their approaches currently?

Thomson Reuters’ Joe Raczynski

Gail Gottehrer:         Sure. They’re a lot of players, so it’s probably all the names that you know. The significant Detroit automakers are all involved. The significant European automakers. Your Toyotas, your GMs, Audi, all those names. An interesting factor has been how technology companies or what’s thought of as more technology companies, the Silicon Valley companies, have become interested and become major developers. Google — its autonomous vehicle arm is called Waymo — is involved. Tesla has been a significant player in this area. Uber is also involved.

It’s interesting, there have been kind of different approaches. Tesla’s approach was that it put its vehicle with technology that it calls autopilot and there was some controversy around this, because some people perceived autopilot meaning that you didn’t have to do anything, but in fact, it was really just kind of an advanced driving system.

It was not full automation, automated driving. It was kind of the cutting edge, the first to be out there on the road, and rather than waiting to perfect the technology, what Tesla did was it had people out driving these cars and they were constantly getting information from these millions of miles being driven and as they gained information about how the system worked, how it reacted to sharp turns, to weather, they would issue over the air updates and do these iterative quick updates to improve the technology and they would push those out to people. The controversial part, which some people took objection to, is essentially it was a beta test on public roads, so it was technology that was learning as people were driving it, but they were driving it on public roads. Essentially, you could be driving next to somebody in this vehicle not knowing that you were participating this beta test.

That was one approach and it’s still going on. The other approach taken by kind of the Googles and more of the traditional automakers was to work on research and development themselves and try to get the technology, do more testing, get it closer to a high confidence level before putting it out on public roads for testing. Rather than doing that iterative learn-as-you-go approach, that worked well for Tesla and continues to, this was kind of a more conservative approach. Then Uber, which I mentioned is kind of an interesting mix. Uber, as a lot of people know, had some issues in California and some other places about whether its drivers were independent contractors or employees and people thought that a ruling that they were employees potentially would ruin Uber’s business model, but thinking ahead Uber was already looking at self-driving ride share. Having the Uber that picks you up, the thinking could be, an autonomous vehicle, so you wouldn’t have a human driver.

Uber was trying this approach and they kind of got into a little bit of hot water themselves last year when they decided that they didn’t want to apply for a testing permit in California. Once you decide that you want to put this vehicle that you want to test on a public road, you have to comply with whatever the requirements are in that state. It is a state-by-state inquiry right now in order to test your vehicle. Uber decided that it didn’t meet the state definition of an autonomous vehicle because it required continuous monitoring by a person. They tried to test without getting the permit and California revoked the registration of those 16 cars because of a lack of a permit, so Uber put those cars on trucks and took them to Arizona where regulation was less stringent. Eventually, they went back to California and applied for the permits and are now testing again in California. Interestingly, now when they test they have two employees in each vehicle for those tests. If you get an Uber right now, it will not be an autonomous Uber.

You’ve been listening to the Legal Executive Institute podcast. Join us again for the second part of our podcast where Joe and Gail will discuss the benefits for law firms in the driverless car technology area.