THE STANDING SENATE COMMITTEE ON TRANSPORT AND COMMUNICATIONS
EVIDENCE
OTTAWA, Wednesday, October 18, 2017
The Standing Senate Committee on Transport and Communications met this day at 6:46 p.m. to study the regulatory and technical issues related to the deployment of connected and automated vehicles.
Senator Michael L. MacDonald (Deputy Chair) in the chair.
The Deputy Chair: This evening the committee will continue its study on connected and automated vehicles. I would like to welcome our witness, Justin Kintz, Senior Director, Americas Policy and Communications at Uber. He is appearing by video conference from Washington.
Mr. Kintz, thank you for attending our meeting. I invite you to begin your presentation, and afterward the senators will have questions.
Justin Kintz, Senior Director, Americas Policy and Communications, Uber: Thank you, chair, and thank you, members of this committee. It is a great honour to be with you tonight. I appreciate your allowing me to report remotely from Washington. I was unable to make it to Ottawa, but I hope to join you next time.
I have a strong affinity for Canada. My maternal grandmother is from Montreal, and I still have cousins everywhere from B.C. to P.E.I. to Toronto, and Montreal, of course, so I’m appreciative and honoured to be here before this august party. I am eager to tell you about self-driving cars; it’s a favourite topic for us at Uber. It is very exciting technology; it will save millions of lives in the future. Uber is very proud to be working in this space because whether or not Uber ends up being the one that defines the future of self-driving technology, the pursuit is noble in its opportunity to save those millions of lives.
I’m appreciative that this committee is taking time to learn about these issues. Of course it’s very early, and this is a good time to study these things.
Let me tell you about Uber and what we’re working on as it relates to self-driving cars. I am responsible for public policy and communications in the Americas, including Canada. I get to work with our tremendous Canadian team that sits everywhere from Montreal to Toronto to Vancouver, and they are learning quite a bit as well in the field of self-driving cars so that they can help educate any stakeholders, public officials, safety officials, anyone in Canada that wants to learn more about self-driving technology.
As a company, Uber is a technology company that uses a smartphone app to connect riders with drivers for private transport. This technology has allowed us to improve mobility and quality of life for people living in and around cities. At first we started in very dense urban environments; and we have expanded since then, over six or seven years, to outlying areas that previously were underserved by other modes of transportation.
We’re seeing tremendous results for people that previously thought they may not be able to access such a service. We’re driving a fundamental shift from the prevailing paradigm of how transportation should be to one of a shared future, and this is a vision of the company: How can we get people to give up the need to have a personal car or own a personal car and get in a shared transport model?
Self-driving is a big part of this. We began our investments in Pittsburgh, Pennsylvania, which is home to some of the finest robotics engineers and AI engineers and self-driving technicians in the world. We opened an R&D facility there a little over three years ago called the Uber Advanced Technologies Group, and we have made significant advances in those short three years in passenger car and trucking technology.
This past May that vision extended north, and we opened our Advanced Technologies Group in Toronto after we partnered with Raquel Urtasun, an associate professor from the University of Toronto, one of the world’s top researchers and top minds in machine learning and artificial intelligence. She leads our Advanced Technologies Group in Toronto, which was our first pursuit outside the United States, and we set up shop at the MaRS Discovery District, which is one of the word’s largest innovation hubs, as I’m sure you all know. We’re pleased to be drawing from some of the world’s top talent in the Toronto-Waterloo Corridor, and we’re very proud to have that presence in Toronto and in Canada.
This commitment was in large part the result of the investments the governments made at both the federal and the provincial level to really invest in the Toronto AI ecosystem in the tech sector. We see ourselves as part of this community and ecosystem. We’ve made significant multi-year financial commitments, to the tune of $5 million as a platinum sponsor of the Vector Institute in Toronto. We’re proud of this investment and would like to grow it. We’re proud of the talent we’re attracting in Toronto.
I want to impress upon you that your efforts for years have been well taken by members of the tech sector like us, and they are starting to bear fruit now. I encourage you to please continue that.
Our investment in self-driving tech has also coincided with our investment in shared-rides technology, what we call uberPOOL, which is a technological way to match riders going to similar destinations on a similar path together so that they can split the fare and do digital carpooling. This has been attempted for decades, but it has not been done well and has not worked at scale well. Uber is trying to now apply our technology in over 40 cities worldwide so that people can have a really affordable trip to where they need to go in a reliable manner, and have a fun social interaction along the way. That’s uberPOOL.
We have that in cities like Toronto, and we believe that cities will become less congested, less polluted, more affordable and more accessible for all sorts of communities when technology like this is allowed to thrive. We are starting to see some of those benefits already.
Self-driving technology is an extension of that and is the next evolution of that. It will bring numerous benefits to Canada by way of road safety, which is probably the most important piece of this. We all know, unfortunately, that driving a car is a dangerous endeavour today. Too often the public has become desensitized to how dangerous it is to be on the roads in Canada or anywhere in the world. We have come to accept that human beings are really great drivers until we’re not. That may mean we are distracted by texting or that someone may be drunk or otherwise impaired in their driving, which can cause a fatal or life-threatening accident.
These are the things that self-driving will immediately impact and improve, but there are other social benefits by way of improved mobility for communities that can’t move around or drive today in an easy away. Of course, there are congestion relief and cleaner air because of the increased efficiency these vehicles will bring. It doesn’t take much more efficiency to take a city like Los Angeles, São Paulo or Mexico City, which today are plagued by bad traffic and air pollution, and take away almost all that traffic if you just increase the efficiency of the system. If we took 10 to 15 per cent of the cars off the road, traffic would largely go away today. That’s very important to us and is something we’re hoping to do with both uberPOOL and self-driving cars.
We understand that while we’re touting safety on the roads and highways, road safety is also probably the biggest concern that the public will have as it relates to self-driving technology for cars and trucks. We lose 1.3 million lives each year globally on the world’s roads to vehicle accidents, and almost 94 per cent of these accidents are caused by human error. That’s something that self-driving vehicles can immediately improve upon. Unlike human drivers, computers are not subject to fatigue or distraction, and they can see in a 360-degree manner in all sorts of different sensory ways, whether infrared or LIDAR — the laser system that spins on the tops of self-driving cars — sonar and other sorts of sensors that help detect the environment around them. It’s certainly a much-improved system from us people who can only see in a limited way.
By perceiving the environment better, calculating different sorts of dynamics faster and reacting earlier, self-driving cars will be able to ultimately drive more safely than people. That’s our goal for the future, and that’s the business incentive for any company looking to get into self-driving cars. The value proposition is how to make this service safer than a human driver. Otherwise, there is no business incentive. Therefore, the incentives are aligned for public or safety officials who are looking at the potential inherent risks and also for the benefits. That’s a really nice alignment.
We know that these are early days. We are still testing these vehicles on the streets of many cities in the U.S. That will expand globally. There are many companies and academic institutions focused on furthering the technology that are testing continuously. We are not yet at the place where we think we can remove drivers from the equation. That may take many years.
There will be important social conversations to be had around when is safe safe enough for self-driving cars, when is the moment when we should consider self-driving cars safer than a person, what does it mean for the future of work — all important questions, and all ones we are thinking today at Uber. We would love to participate in any discussions that the Parliament, the Senate and this committee are having with regard to these issues.
I really appreciate having the opportunity tonight to give a brief summary of what we’re working on. I welcome any questions you might have. Thank you so much.
The Deputy Chair: Thank you for your presentation. We’ll start with questions.
Senator Griffin: Thanks for your presentation. I have a couple of articles; I don’t know if you are familiar with them. One is from Ottawa, and this appeared on Twitter, I think. It’s called “Uber Assist, a ride service for people with disabilities, launches in Ottawa.” I don’t know if you are familiar with that. It’s the program uberASSIST. Could you tell me a bit more about it?
I’m from Prince Edward Island. We have something called Pat and the Elephant, a fleet of little vans that can take people with disabilities. How would you see this comparing with something like that?
Mr. Kintz: Thank you, Senator Griffin. I appreciate your concern for this area. It’s something Uber has thought a lot about since our early days. uberASSIST is progress toward an accessible future for ride-sharing technology. It’s not all the way where we want it to be. It’s basically an educational tool that allows drivers who are specially trained to help people who may be disabled or elderly and need a little extra care and help getting in the vehicle and people who may have a wheelchair that can be collapsed and put into the trunk to use services like Uber to get where they need to go. They can select that option in Ottawa and other places where we have uberASSIST, and they will know that they will get a driver who is familiar with how to properly stow a wheelchair and deal with the needs they may have.
There is also a whole range of other folks with disabilities who need a little extra attention or who might need special hardware by way of a wheelchair-accessible vehicle, something bigger that can fit an electric wheelchair and things like that. Those are things we are focused on now: How can we as a digital platform make it economical for the providers —the drivers — to retrofit their vehicles or to buy new wheelchair-accessible vehicles, and get the service to a place where there are enough vehicles online at any given time, whether through our platform or perhaps in conjunction with the rest of the industry, so that someone who needs a wheelchair-accessible vehicle can push a button and know they will be able to get a ride in a short amount of time, reliably?
Today, we could probably provide enough of a service in some cities so that people can get a ride from one place to another, but we need to have enough density in the system, with enough drivers with proper vehicles, so that you can get to where you’re getting going and also get a ride home in the same manner.
We are working with lots of experts in this field to try to figure out the best way to do that, the best ways to incentivize drivers to make those investments, and we’re looking at the best ways to work with governments to form public-private partnerships, perhaps to set up a funding pool to incentivize drivers to make those kinds of investments.
We’re making progress in a lot of different cities in this regard, and, senator, I’m eager to share with you some of our findings because perhaps it could be interesting for places like Prince Edward Island, but it’s certainly something that transport officials have struggled with for decades. Uber has found it a very complex challenge as well, but we have the size, scale and resources now, I think, to tackle this in a way that hasn’t before been seen, and it’s our responsibility to do so as a transportation and technology service that can have that capability. It’s just the right thing to do.
To your question about self-driving cars, I think all of that applies to self-driving cars as well. We’re going to need accessible options for self-driving because the folks who will benefit most from the service are members of the disabled community and people who are visually impaired who can’t just hop behind the wheel of a car and drive themselves somewhere. They have to request a ride that might take up to two hours to arrive. I think self-driving cars will be transformative for those people.
Senator Griffin: Thank you. You have made the point about cities, of course, and that’s where there’s a critical mass of people. How does Uber work in the countryside and in the rural areas where the population is more sparse? Is it very active in rural and country areas?
Mr. Kintz: It’s less active, but we’re growing, and I think it’s important to look at the history of how Uber starts in a particular area. When we first started, we only had a black-car service, and there are only so many places that have enough black cars to make that worthwhile; it’s only the big cities. As we grew, we realized we could have a low-cost option, UberX, which we call ride sharing.
With ride sharing, you don’t need a Lincoln Town Car and you don’t need to be a commercial black-car driver to operate in a place. You can actually be me with my Ford Escape, and you can have access to the platform as long as you can pass a number of hurdles like background checks, vehicle inspections, insurance requirements and other consumer protection hurdles. Then you can go online and take your neighbours where they need to go, and in exchange you can make a little money in a flexible way. And you can do that for five hours a week or 35 hours a week if you wish.
That has opened up a lot of opportunities to us because we’ve been able to see that as that system grows, the demand grows with it along the edges. So we used to start these services in dense urban areas like the central business district of a city, and then we could see people opening the app and you could see that demand grow along the edges because people knew about the service and wanted it themselves. So the service coverage area would expand, going from the central business district to the whole metropolitan area and then beyond, in some cases. Today you see people even commuting long distances using UberX. That’s true as well in rural areas.
Part of the problem is you want to make sure the service is reliable no matter who is using it. If I live in the countryside and at 8 p.m. my wife and I want to go on a date night to a restaurant, it doesn’t do anyone much good if I can get to a restaurant but can’t get home. You have to make sure there are enough drivers online at all times so that there is reliability in the system. We won’t put the product out if that reliability doesn’t exist.
Most of the time it just takes a little time for enough demand and awareness to build in a given place where you get enough drivers to sign up thinking it could be an interesting opportunity. We’re accessing areas now that I never expected to be able to access. It may not be a service available everywhere all the time, but that is our vision. We’d like to get to that.
Senator Eggleton: As a senator from Toronto and a former mayor of the city, I’m pleased to know that you have set up shop in Toronto, particularly at the MaRS centre. Can you talk a little bit about the specific activities that you’re planning in Toronto in this connection with automated and connected vehicles?
Mr. Kintz: Thank you, senator. For the time being, it’s a lot of research and development, and it’s focusing, at least in the case of our Advanced Technologies Group in Toronto, on how to apply machine learning to the scenarios that the self-driving vehicles will need to learn.
As an example of this, if you have a 16-year-old who is learning how to drive, they’re not a very good driver, from most people’s experience. Maybe there are some rare ones out there that are natural drivers, but most people aren’t great drivers at age 16. By the time you put in a million miles of driving, maybe that takes 30 or 40 years, but you’re a pretty good and safe driver, which is why the insurance companies tend to apply higher premiums to young people, because they don’t have that experience.
Self-driving cars are interesting in that you don’t necessarily need to have each vehicle learn that million miles by itself before it becomes a safe driver. In fact, you could have, let’s say, 1,000 vehicles, and each one is out experiencing different things in its own environment; it might encounter something new every day, like a new pothole here, a new traffic light there or a paper bag that blows in the road.
All those learnings are brought into our data facilities, where they are analyzed and learned from, and then we can upload all those learnings to the rest of the fleet so that the other thousand cars are learning instantly, along with that one car that spotted the paper bag, what that object is, how to perceive it and whether it’s safe or something they need to be wary of. Then the next time out they have all that learning built into them, so it’s an incredible way to learn at a compounded rate.
The machine learning helps to accelerate that because it can set up simulations, do different scenarios and run them at a rate faster than the human brain can operate. It can help accelerate that speed of learning for the self-driving cars, making them safer by the moment. These vehicles are taking in a tonne of data, and that’s why a lot of real-world, rich data is necessary so that we can collect that and put it back into the systems, and they can learn faster. It all builds on itself, and that’s very important.
That’s why you see Uber and a lot of our competitors relying heavily on figuring out how many ways we can get out and get real road miles for passenger cars and vehicles.
For the sake of Toronto, that will be the focus of the R&D facility for some time. When we get to the moment when we are able to deploy self-driving vehicles as a shared fleet within the Uber network, naturally Toronto will be a very appealing destination to have that technology, and it will be a technology that gets mixed in with the existing UberX network. We have no plans at this time to expand anywhere beyond our current testing in Pittsburgh and in Tempe, Arizona.
That’s in part because, in Toronto’s case, self-driving cars are much like Californians: They don’t love the snow. So we’re working on figuring out ways to make the sensors that go into the optics of the vehicle less susceptible to disruption by snow, dust or sunlight coming in at twilight that might affect the sensors. A lot of technology needs to be improved.
Senator Eggleton: Thank you for that detail. Your vision is based on ride sharing. I don’t know how much of that you can do prior to seeing a Level 5 automated vehicle, but I guess you do some of that now. But the fully automated vehicle should mean less traffic and parking if it’s a shared service. It can spend a lot more time in motion. But that requires a cultural change because people are used to owning their own vehicles.
How are you going to get people out of this cultural circumstance that exists now — that is, people owning their own cars — to getting into more of a ride-sharing kind of vision that fits with what you want to do?
Mr. Kintz: That’s a really good question.
That is important not only for the sake of the shared-ride future vision but also, in the shorter term, for the sake of public acceptance of the technology and the public’s comfortableness with what otherwise could be perceived as a scary thing. That is, no one is driving these headless vehicles. What will we do? That’s something we need to be aware of. You can talk about it all day long, but the only way to do that is to make sure that the proof is in the pudding — in this case, that the product works reliably, is safer and performs better and smoother than a human driver could. Until that point I don’t think the public will trust it, which is fine. That gives us the requirement that we need to make sure the technology is there and convincing to the public.
I’ve gotten to see a bit of this as we started to do public city street testing in Pittsburgh where today you can pick up someone in an Uber X that’s a self-driving vehicle. There is still a safety engineer behind the wheel to intercede if the system becomes unconfident, but it’s a shocker to people on the sidewalk and passersby who see this self-driving car going down the street. It’s changed a lot over the last year since we launched in August 2016. Then you’d be in a self-driving Uber in Pittsburgh and see tourists and residents walking on the sidewalk standing there with their jaws on the sidewalk, taking out their phones and taking a snapchat or something of the vehicle. However, if you go in one today, nobody looks up from their phone. They are still texting furiously as they walk down the street. The public is becoming desensitized to the technology in an important way. Of course, we have to talk about the statistics of potential safety gains, but most of the public is not yet familiar with this technology. It will take time for the public to really appreciate it. It will take some interaction with the technology. That’s why we think it’s important to integrate it today with our services where we can to make sure people get a chance to get in the seat of a self-driving car and see it for themselves.
Senator Eggleton: I could ask you questions about cybersecurity and things like that, but I think my colleagues will get into that.
I’m interested in the fact that you conducted a pilot project in Gainsville, Florida, involving seniors, a local elder care network. I don’t know how automated or connected the vehicles were that you got them into, but were there any lessons learned from that pilot project that might be valuable to the acceptance of older folks with AV and CV technology?
Mr. Kintz: That’s a great question. Thank you, senator. I went to school in Gainsville, Florida, so this is near and dear to my heart. I think we learned a few things. Chiefly, I was reminded politely by the AARP representatives when I approached them years ago to say it might be interesting but do enough seniors or retirees use smartphones to allow them to use the service? They said, “Excuse me? More seniors carry smartphones than any other demographic in the population.” They put me in my place on that, and it’s true. What we have seen from partnering with AARP and other organizations that work with seniors and retirees is that no matter where you go in the world, this part of the population desperately needs mobility. It’s freeing and independent for them at a personal level. In many cases it’s important at a health level to make sure they can get to their doctor appointments and go to the pharmacy to get their prescriptions, get home and to the grocery store, and so on.
This is personal to me. My own father has Parkinson’s disease, and it’s becoming challenging for him to drive. I order him Ubers as frequently as possible. We have learned that a lot of kids do order Uber for their parents, or medical providers do. There were complaints coming through the system, for example, “I can’t have Uber while my mother is taking a trip, but I would like to be able to do both — that is, order something for a senior in need and at the same time go on a trip myself.” So we developed technology to do that, namely, where you can send an Uber to someone you like. We also developed a centralized portal like the one being used in Gainsville, where retiree centres or living facilities can order vehicles from a centralized dispatch for patients or for residents to be able to go where they need to go. That’s very important, and it’s growing.
The other thing we learned is that there’s an interest in peer-to-peer service. This is something that AARP raised with me as well. We have launched a similar service in a very aged community in Japan, where almost the whole population is 60-plus. It’s a peer-to-peer service where seniors are driving seniors. There’s a lot of trust involved there. There are a lot of commonalities on a social level, and it creates a great experience. That’s not scalable everywhere, but it’s an interesting and fun service that we’re able to provide in Japan. These are things we’re thinking about, senator, so you struck the right note.
Senator Bovey: Thank you for your presentation. Some of my questions have been asked, so I will move in a different direction, but first one more with disabled users.
We’ve met with representatives of several groups who haven’t, I guess, maybe thought through the technology as much as they’re going to in the future and who have been looking at it with its challenges, shall we say. I wonder what kind of work you’ve done one on one with various organizations of people who are not just a parent here and there but disabled groups across the country to get them thinking about what the technology could mean for them.
Mr. Kintz: To make sure I’m on the right track, you mean in terms of Uber as a general service, not just self-driving cars?
Senator Bovey: I’m thinking of the shared driving experiences.
Mr. Kintz: Great.
Senator Bovey: Not just the fear of the self-driving vehicle but the fear of being with people they may not know or the fear of making sure there are people to help them. You addressed the equipment a bit. Also fear of how they can pay. Some of the questions they raised were very small but big, if you know what I mean. I guess the fear of not knowing who’s in the car with them comes with a greater sense of vulnerability perhaps than for those without disabilities, which gets into the question of data protection and privacy protection. I’m moving from one aspect into the larger aspect.
Mr. Kintz: Okay. Thank you, senator. First, it’s important that I admit we’re far from having great answers on all these things, really. It’s incumbent on us to make sure we’re at least providing incremental improvements along the way.
If you take just the accessibility piece, I think many groups see us as an interesting opportunity as we’ve gotten so big and have so many resources available to us. They say, “You’ve come up with an interesting service here that’s fantastic for a lot of different groups of people, but what about us? We’ve been talking about this sort of thing for decades. Yet here you are. This is it. This is the opportunity we’ve been waiting for, and you’re not taking seriously the needs of our community.”
I would say that at least for the early years at Uber, they were absolutely right. We didn’t appreciate the gains that could be made in serving those kinds of communities because we were so focused on growing the service globally so quickly that we didn’t stop to assess how we could refine the technology and get it to a place that can work well for communities that aren’t just the general population of people using the app.
Their criticisms were spot on. In many cases, once we stopped and listened, we started to learn a lot. I think that sitting down and taking time to listen and have a dialogue was something we didn’t appreciate before. We do now and we have for the last couple of years. We’re learning a lot about how we can do these things better.
One of the warnings was this: The disabled community is not homogenous. Of course, wheelchair-accessible vehicles is the biggest challenge from a transport perspective, but it turned out there we could do some really expedited fixes to the app that really help accommodate visually impaired users on the rider side, of course, and it’s a software update that many software providers use to allow the app to have sort of a voice feedback.
Then for the hearing impaired, both on the rider side and that driver side, this was a really exciting project we got to work on where we were able to update the app to allow deaf or hearing-impaired drivers to communicate properly with riders through the app. It just creates a two-way communication system that says to a rider — once you get the request, if you’ve taken Uber frequently, you may have run into this yourselves — it says, “Attention, your driver is hearing impaired, so we ask that you please communicate through this interface so that they know where you are, where you’re going” and that sort of thing. It has created a lot of opportunities, both on the income-earning side for drivers who may otherwise have had some limitations on what they could do as a profession, and on the rider side it’s a relief for a lot of people who otherwise would have a really difficult time interfacing with a transportation provider who may not know they have a hearing impairment. That’s really exciting for us. We’ve seen amazing stories come out of that.
To your other point about people’s privacy, we’re looking at now partnering with medical transport agencies, veterans hospitals and other sorts of hospitals and medical providers. I will speak for the moment to the U.S. system where we’ve been having active conversations with the federal health and human services division and the state divisions as well to make sure that we’re able to protect any sort of patient privacy while at the same time providing some forms of non-emergency medical transport. We think we could bring a huge amount of cost savings and efficiency to a system that has suffered in regard to both of those things for a very long time. The tricky part is making sure that you respect those patient rights.
I would love to have our health team follow up with you, senator, if you’d like to learn more about that, because I don’t have a tonne of the specifics myself at the moment, but I do want to make sure you and the committee know we’re thinking a lot about these issues because we’ve been hearing it from consumers. It’s something riders and drivers want, so we’re committed to fixing them.
Senator Bovey: I think the data protection is a very important one, and it’s the link into cybersecurity as well, which my senator colleague brought up a minute ago. I don’t know if you’re ready to venture into the field of questions about cybersafety.
Mr. Kintz: Yes, absolutely. It’s a relevant topic that’s brought up frequently, especially as it relates to self-driving cars. I’ll try to focus on that.
Cybersecurity is not unique to self-driving vehicles or to software companies. In fact, we have, I think, over a billion vehicles on the road today that are normal passenger cars with vulnerabilities as they are starting to get these more advanced software functions in each one. We’ve seen this result in some form or another of different sorts of breaches of their systems, so I would propose to this committee that we actually have a more pressing concern today with the cybersecurity vulnerabilities of existing automobiles that are on the road that are not self-driving cars or anything special. That’s something to consider. In fact, the infamous Jeep hackers who were able to take some form of control over a Jeep a couple of years ago were two people we hired to lead our cybersecurity defence team for the Advanced Technologies Group. They helped us build a cybersecurity defence for self-driving vehicles that I think is very advanced and very sophisticated. It’s important for the committee to know that self-driving vehicles are localized, so often these arguments are thrown out that say, “What if a hacker from Russia takes over an entire fleet of vehicles? There could be a threat of national security. You could drive someone off a bridge.” I suppose nothing is impossible, but I do think that is a bit overblown in that, at the moment, you can really only take over one vehicle at a time.
Also, it’s important to note that the companies involved in self-driving vehicles are advanced software development professionals. Cybersecurity as a whole is a cat-and-mouse game, and that’s true whether you’re a software company, an automotive company or a government. You have to make sure you are always developing defences that are one step ahead of the bad guys. In this case, with self-driving vehicles, we feel very confident that we’ve recruited some of the world’s finest minds as relates to cybersecurity.
We regularly hold bug bounty programs or we will invite hackers to try to breach our defences to see where the vulnerabilities are, so we take it very seriously. To date, that’s really the industry standard. I don’t think there’s anyone who can rightfully get up before this committee and say, “We’re completely invulnerable.” It’s just not the case for any software company or organization worldwide, but we take it very seriously.
Senator Boisvenu: Good afternoon, Mr. Kintz. As your origins are from Montreal, what about practising your French tonight?
Mr. Kintz: That would not go well.
[Translation]
Senator Boisvenu: I will continue along the same lines as my colleague, Senator Eggleton, with regard to your centre in Toronto. So it will be a research centre, as you said.
My first question pertains to Canadian infrastructure. We know that this type of vehicle will require specific infrastructure. I think you can see what is being done in other countries, in other provinces. How would you compare Canadian infrastructure to that in other countries in terms of advancement or degree of preparedness for this new technology?
[English]
Mr. Kintz: Thank you, senator. It’s a very good question. It’s one that governments worldwide are asking of our industry and asking of themselves.
I think that, first of all, the infrastructure in Canada is well suited to welcome in testing, deployment and proliferation of these sorts of vehicles. I think that for those who are very invested in building out a lot of smart infrastructure and next generational infrastructure, my answer may be disappointing, which is that we’re working very hard on our end to make sure the technology can work. I mentioned it’s not great in the snow. We would like to get to the point where it works great in all kinds of weather and infrastructure, such as bad roads, bridges, all sorts of challenging environments. So we want to achieve that level of technological competence so that the trip is safe no matter what the vehicle might encounter.
Now, there are other interests in this space. Sometimes automotive makers or others who want to get involved in self-driving are interested in vehicle-to-vehicle communications or vehicle-to-infrastructure communications. I would say that at a minimum, self-driving vehicles will work best when you have the basics done well, which are roads that are well paved, signs that are clearly marked and visible, traffic lights that work properly and are synchronized and lane stripes that are painted properly.
Our self-driving trucks work a bit differently from our self-driving passenger cars in that they’re not really designed to be able to handle complex urban environments like the passenger cars. The self-driving trucks are designed to be able to handle straight-ahead highway miles, and the technology relies on the lane stripes so that it can create virtual bumpers for itself to keep it in the proper lane. Much like if you used the Tesla today on their autopilot system — it’s using the lane stripes to keep it in the right place.
When the lane stripes begin to fade, the system loses its confidence and won’t work properly. These basics are what we’re interested in at the moment. The other sorts of investments in smart traffic lights and road signs are very expensive for any nation’s infrastructure. They can be used interestingly and would make the whole ecosystem safer, so I don’t want to discourage that. There may be some interests that are looking to build those up that would help everyone and ultimately make it safer, so we encourage this. It’s not something we need at the moment for our purposes, but it may be something that other makers of this technology would like to have.
Finally, I would say the infrastructure around making sure that we know where construction is happening, where big events are happening, the conversation and relationship with law enforcement or with local transport officials is a very important one because while self-driving cars can map out an area and know where the signs and traffic lights are, what the other cars will be doing, it may not know where overnight construction has begun. And that may change the dynamics for that system and how it is able to perceive the world. That is why having that relationship is important. No matter where the regulation of self-driving ends up happening, whether it’s at the federal, provincial or local level, it will be important that any company has a very good relationship with the local government so that we’re able to have that dialogue in real time.
[Translation]
Senator Boisvenu: We see that smart car technology is being developed in various parts of the world and in North America, so there are a lot of researchers and manufacturers. Will all these technologies compete with each other or will they complement each other? Moreover, does this represent a challenge in terms of the infrastructure that will be needed for these technologies, whether they are shared or disparate?
[English]
Mr. Kintz: We believe that shared is the most likely future for this technology. I think the automakers — and I don’t mean to speak for them — would also say that they think shared fleets for the passenger car side are the future model. For commercial trucking, it’s already a fleet-wide model, so it’s a more natural step for them.
You see a lot of automakers who are finding it unlikely that there will be a future where they will be designing consumer products that will do all the things in a self-driving manner that someone might need, no matter where they want to go; it’s unlikely with self-driving cars. It’s very time-consuming and intensive work to properly map an area so that self-driving cars can perceive their environment properly. If you think about Google maps or some sort of traffic mapping technology, you need to make sure you get the streets and the stop signs right, where those are, and the names of the roads, and that can be accurate to a matter of metres. But with self-driving car mapping, you want to be accurate to a matter of centimetres because those metres might mean the difference between proceeding with the flow of traffic or veering into another lane and rushing into oncoming traffic, which can be dangerous. So the mapping needs to be precise, and you have to spend a lot of time making sure you’re mapping it over and over again.
And you could have other things happen. The spring could emerge and you could have a bush grow in front of a stop sign. It’s very important that you have consistent mapping so that you know there is a stop sign there, and so that the vehicle sensors wouldn’t miss something like that. I think it’s unlikely that you would have that level of mapping, at least until the technology evolves to a place where the mapping can be done quickly and efficiently. It will be unlikely you can have a self-driving car — a Level 5 — that can go anywhere, do anything on its own without any sort of operator control.
I think it’s much more likely and more efficient for transport systems around the world to have shared self-driving fleets that can be operating 24 hours a day and be picking up riders in a pooled scenario, like uberPOOL, where you have two, three or four riders at a time.
The vehicles might be designed in a more interesting way when you don’t have to contemplate having a steering wheel, pedals and other motor vehicle safety standards. You can have the seats pointed in at each other, you can have a conversation, play checkers, watch television with entertainment systems inside. You could do a lot of things when you don’t have to worry about the driving controls.
So you will start to see the auto manufacturers producing those vehicles. You can have all sorts of consumer products meant to be used in a shared manner, different sorts of pods featuring different environments. Maybe you go into one and it’s only showing the day’s news, or another one might be showing cartoons for the kids and another might have sports on the screen, and you can choose. So the future is limitless when you think about what this might turn into, but it’s important to see the trend, which is the fleet-wide model.
And that’s the one that will also result in the most efficient and socially beneficial outcomes. Because then you’re talking about reducing parking, congestion, air pollution, all the things that plague some of the world’s big cities today. It’s not just something that will be nice to have. It will be necessary. If you look at the world’s great cities in this trend of urbanization, some cities are reaching the tipping point where it’s becoming almost unlivable because the pollution or congestion is so bad.
I frequently work in Mexico City and in Rio and São Paulo where it takes three or four hours to go 20 or 30 kilometres. It’s unsustainable, and that’s something we all want to work towards.
Senator Mercer: Thank you for your presentation, Mr. Kintz.
I want to take you out of the city for a moment and put you on the highway with some trucks, in a truck train or whatever we’re going to end up calling that when we have a number of trucks going along with little supervision.
This tool requires proper maintenance of lines on the highway signs and the highway so that it will be safe, and I see this as probably the biggest opportunity here. In this country, by 2020 we’ll have a shortage of over 33,000 long-haul truck drivers, and in a country as big as ours, long-haul truckers are important, and we will be down by 33,000.
So this is a possible answer to our problem.
But then I ask this question: What does the Teamsters union say about this?
Mr. Kintz: Thank you, Senator Mercer, for hitting on one of the more controversial parts of this technology. The Teamsters have already started to make noise about this in Washington. The U.S. Congress is taking up legislation that would lay out the first rules of the road in the U.S. at a federal level for self-driving cars.
They tried to include self-driving trucks, and the Teamsters union and other advocates stood in the way of that. I would say that your analysis is spot on, which is that there is currently a shortage of truck drivers. The technology, as designed today, is intended to be something more akin to autopilot on an airplane; it helps improve the safety of the truck drivers on the road, because much like passenger cars, they suffer from really a high rate of fatal road accidents. In fact, it’s disproportionally high compared to passenger vehicles, and these heavy vehicles are a danger to everyone else on the road as well.
So the self-driving truck technology would allow a truck driver to be able to step away from the wheel, maybe rest in the cab, maybe get some sleep, work on paperwork, call home, whatever he needs to do. When the truck then reaches its destination, the technology would disengage once it hits the ramp, and the driver would need to take back the vehicle to the loading dock. At that point it would certainly require human intercession to make sure it gets to the loading dock safely and the cargo gets unloaded.
We found that is probably the biggest value proposition for the trucking industry, which is making sure that those highway miles are covered by a safer technology like this that could also be very convenient.
It would be extraordinarily complex to try to figure out how to get from the ramp to the loading dock and do the unloading, so we contemplate a future where the driver, or at least some sort of a helper, is involved. I think the labour unions will say what they have to say about this in the U.S. and Canada and anywhere else, but I think what they’re missing, in my mind, is the extreme safety benefits that could be realized for truckers.
It’s not a very healthy profession. In many cases truckers suffer, in addition to the road safety, from health detriments from being on the road and behind the wheel so frequently. It could help them get more rest.
Also, when you apply the self-driving technology to our other product, called Uber Freight, which is basically a digital brokerage that links the suppliers to the drivers and allows them to find loads they can take home or on a route that they would like to go, you realize it’s potentially a way for the truckers to make more money more efficiently because they don’t have to deadhead on the way back. That could mean more time at home with the family and less time on the road.
I see a lot of potential benefits for the drivers. We do interact frequently with trucking associations and groups of drivers that are already seeing the benefits of Uber Freight, where we launched it in Texas and in some places in California. We’re excited about the pairing of those two systems, self-driving technology mixed with the Uber Freight brokerage technology.
Senator Mercer: I want to go back to something you said earlier. You talked about snow. Snow happens, not just in Canada but in a large part of the United States as well. To make this technology work at all, particularly for trucking, you’re going to have to tackle the problem of snow. I drove down through the Blue Ridge Mountains a number of years ago, and I came out on the highway early in the morning where it had snowed overnight. It wasn’t much snow for me, but there was enough snow that it caused problems for everyone else. I thought everyone was going slow when I came up on a hill. They weren’t going slow. They were spinning their wheels. I kept moving by because I knew what I was doing and they didn’t. Most of the vehicles that I was driving by were big transport trucks. They weren’t prepared to handle this small amount of snow.
How will you get around this technology in communities where it doesn’t happen very often? I was going down to North Carolina where there is not a lot of snow and through West Virginia where it does snow a bit and suddenly you go from one place to another. You are driving along in an area where there is no snow and suddenly, unexpectedly in this time of climate change where we have strange weather systems, you have to deal with snow. You cannot put vehicles on the road that are not adaptable to those changing weather conditions.
Mr. Kintz: You’re absolutely right, senator. The good news is that the technology is advancing quickly. This is more dependent on the sensory technology advancing quickly, so that the high-definition cameras that are used are able to deal with things like dust storms and snow and poor visibility. We’ve seen an exponential increase in the advancement of that technology at a very low cost. It’s almost cliché to say, but it’s only a matter of time before that technology advances. We’re probably talking only in terms of years there. We’re not talking decades.
Most of our competitors have been testing and start out in California, and we have similar testing happening in Phoenix, Arizona, but we also have our R&D facility in Pittsburgh. They are no stranger to snow, bad weather, poor road conditions in Pittsburgh. In fact, our engineers decided to begin our first deployment there because they wanted to test the sensors and the software out in those kinds of conditions. In fact, I think Pittsburgh has the most bridges of any city in the United States, and bridges are very difficult for the sensors as well.
We thought that if we’re going to learn in a quick manner, let’s test it in what the engineers like to call the double black diamond of self-driving cars, Pittsburgh, where you have bridges, snow and in some cases outdated road infrastructure.
Senator Mercer: And there are a lot of hills in Pittsburgh, too.
Mr. Kintz: Exactly right. In fact, I’ve taken rides where the snow has piled up. I had a delegation from the UAE in Pittsburgh last winter. We did a trip. I thought that the snow would pile up all over the sensors and wondered how this would go. These folks from the UAE certainly are not going to care about the snow, so I hoped it would go well, and it went really well. You do see incremental improvements happening in dealing with the weather. It will only be a matter of a couple of years before the sensors are able to handle almost any sort of weather conditions.
One more note on that. There are also incredible advantages that we can’t realize as people driving, such as in the middle of the night encountering a deer crossing the road. You and I see that as soon as it hits our headlights. In fact, that could sometimes be too late, whether for the deer or ourselves in a dangerous way, but the LIDAR systems attached to self-driving cars could spot that deer from 100 metres away in the middle of the night and respond safely without you noticing a hitch, without hitting the brakes very hard. There are a bunch of interesting advantages that we can’t even contemplate today based on our human experience.
Senator Bovey: I have one question on this round. As you know, we’re a committee of the Senate of Canada, and part of our job is to make recommendations to the federal government as to the kind of legislation and regulations. I would be interested in your advice as to what kinds of regulations we should be looking at as a federal government, when driver licencing and roadways are provincial and municipal responsibilities. I would be interested in knowing your regulatory thoughts.
Mr. Kintz: Thank you, senator. I receive this question frequently from governments all over the world and across the U.S. at the federal, state and local levels. I think it’s important to study this issue, as you’re doing tonight, because it is very early in the lifespan of this technology. There is plenty of time to lay out thoughtful rules of the road, and I think that further study is certainly warranted and advised here because the technology is going to evolve very quickly from here and very often. So I fear that if any sort of comprehensive rules were prescribed today, they would be hopelessly out of date within a year, and in some cases could even create a chilling effect on the technology that could delay the safety benefits for years or even a decade.
I think what we’ve seen in places like, let’s say, the U.S. Congress at the moment, they are looking to try to expand the exemptions from federal motor vehicle safety standards for the automakers so that they can start experimenting with vehicles that don’t necessarily contemplate having a person behind the wheel. It’s those cars I mentioned that could be entertainment centres or social centres where the seats are facing one another. That’s something that the manufacturers are interested in starting to do, so the federal government is acting on that front.
You have seen a lot of the U.S. states start to address auto liability concerns from the insurance industry and the trial bar. That is going to be an evolving discussion that’s probably taken up annually by those groups, and they’re starting to try to get their heads around what this might mean for personal auto insurance lines versus commercial auto insurance lines. They are trying to get ahead of the curve for the industry. It may make sense for them to prepare for that and to adapt the laws accordingly.
I think the other laws that are being written today, whether it’s in the U.S., Australia, New Zealand, Singapore, Dubai or other places that are looking at this very closely, you’re seeing maybe some struggle to find how safe is safe enough. How safe is a person compared to how safe is a self-driving car? I don’t yet have the answer for you. We plan to contribute to that discussion by submitting some of our findings and data as they relates to safety.
You saw our competitor Waymo that a couple of days ago released their own safety report — the first of its kind — that helps detail some of the findings. I hope others in the industry follow with that. I think we could all learn a lot. I think there would be a very short-term play for any company or organization that’s working on this technology to try to hide the ball from any government in a meaningful way, because we can all learn from each other’s progress on the safety front.
That said, there is also sensitive and valuable competitive and proprietary data that we, as companies, are looking to protect, so there has to be a fine line. But at the moment, we’re very eager to work together with anyone who is thinking about this from a governmental level to figure out what the rules of the road should be. Luckily, we have the luxury of time to make sure that those rules are set up well to help embrace and encourage the technology to get to a safe level while not stifling it.
I would leave you with one final thought: It’s most important that if there are standards to be set now, I would encourage that we focus on the hardware requirements, and that could be the vehicle itself — that’s already a regulated industry, already something that has been carefully studied — or perhaps on the standards around the optics, whether it be how far the lasers should go in the LIDAR system, 100 or 120 metres, and whether we should set some standards there. Should the cameras be able to see to a certain definition and that sort of thing, so that the whole industry can adapt to it.
I would caution that it would probably be premature to try to set in place any sort of restrictions around the development of the software systems which are updated multiple times daily. We’ll need to go through probably thousands of iterations before we get to a scaled-down consumer service. It would be very stifling to that technology if a governmental body or some sort of third party needed to approve the evolution of that technology from a software perspective along the way. That’s my only note of caution.
Senator Bovey: Thank you very much for that.
The Deputy Chair: Mr. Kintz, on behalf of the Senate of Canada, I would like to thank you for your participation today. We very much appreciate you taking the time to speak to us.
(The committee continued in camera.)
The Deputy Chair: Senators, is it agreed that the budget application for the special study on the regulatory and technical issues related to the deployment of connected and automated vehicles for the fiscal year ending March 31, 2018, in the amount of $47,145, be approved for submission to the Standing Committee on Internal Economy, Budgets and Administration?
Hon. Senators: Agreed.
The Deputy Chair: Thank you, honourable senators, for your patience. For our next meeting, next Tuesday, we will hear from witnesses from the Central North American Trade Corridor Association.
(The committee adjourned.)