Among the plethora of automotive technologies unveiled by car makers at this year’s CES, the integration of AI was easily the most noteworthy, and it’s easy to see why. The integration of AI in cars, which includes safer driving, less traffic congestion and eventually, full automation, has the potential to revolutionize the way think about transportation. However, the transition from having cars with Siri-like assistants to fully-automated vehicles is not set in stone, and there are many unanswered questions as to how we’ll actually get there.
To shed light on this fascinating subject, we teamed up with Toyota – makers of the Concept-i automobile that was unveiled at CES 2017 – to speak with three industry experts: Sidney D’Mello, an assistant professor of psychology and computer science at Notre Dame University, Dan Sturges, an adjunct professor of transportation at the College for Creative Studies in Detroit and Fei-Fei Li, the director of Stanford’s AI Lab.
Q: How do you see the future of cars and transportation unfolding in the coming years?
Dan S: The future is not just product design; it’s transportation planning, urban design, all of these things together. I’ve spent most of my adult life trying to right-size transportation. For example in Denver, Colorado, the average person in a car is 1.1 people per car. You’ve got four empty seats. It’s a 20% load factor. If you’re an airline company, you can only survive a week with that load factor.
That’s an issue that affects congestion for sure, but it’s also an efficiency thing, both in terms of energy use as well as cost of transportation for the consumer. You hear a lot about it now with the self-driving car as a way to lower the cost and get more people moving.
Sidney: There’s more of a potential for machines, obviously, now to communicate. Communication is the big key. For example, all the traditional communication is still the user giving instructions to the machine and that really hasn’t changed much. In the 70’s, they were typing commands, and in the 90’s you’re moving your mouse and again, selecting commands. Now you have swiping and gestures and it’s more natural, but it’s still the same thing, you’re instructing the machine.
This is completely different from how humans communicate. There’s less instruction going on. It’s more of a conversation and you communicate through words, through intentions, and through feelings. We’re moving towards that in some ways. That’s promising.
Q: How are advanced technologies and automotive improvements going to affect safety for drivers and the public in general?
Dan S: Right now we have people with all of these different travel behaviors buying a similar car. The car is trying to be safe for all of these different trips. If we start to go into a more right-sized model with maybe smaller transportation in a community, that’s going to be safer for everybody around it. People walking, people biking, kids and everybody are going to be safer around smaller vehicles in these environments.
When you need to travel far, that’s a specific trip. That car could have the exact kind of safety equipment I need, and it’s incredibly well designed for that type of use. We’re going to move into a more customized and context sensitive type of a transportation future.
Fei Fei: I think it’s not just driver, it should be safety of everyone. Driver safety can not be at the expense of passenger, pedestrian, and anything else. I think that is a key question is that we believe self-driving car, because of it’s much better potential to be safer and more efficient, humans are extremely limited in our own biological construct in terms of attention, motor skills, perception, and all those.
I’m optimistic in terms of making safer through self-driving cars. But I also want to emphasize that this is still a long road. Companies should be doing it responsibly, not just from a technology point of view but from a society point of view. Thinking deeply about the people, the fabric of our society. The people, drivers, the passengers, the families, the workers, the labor force. These are all important things to consider.
Sidney: There’s a fundamental barrier to cross, and that’s attention. There’s more and more work happening in eye tracking now. More and more of those eye-tracking devices are coming out of the labs and they’re actually getting into the world.
One example is reading: You’re reading something dense and you just zoned out and you don’t even realize, as a human, that you’ve been zoning out until you catch yourself. Many railway accidents have been attributed to somebody not paying attention to speed limits. There’s lot of potential in that realm. We don’t even actually react immediately. You could be staring at the page — or the road — but processing nothing.
This means that there’s a possibility of the machine to be one step ahead of the human in terms of knowing their own feelings and mental states.
Q: How do you see our relationship with vehicles evolving as Artificial Intelligence, autonomous driving, and ride sharing become more advanced and ubiquitous?
Fei Fei: Humanity has entered into an era that we need to think about co-working, cohabiting, coexisting with intelligent machines. This might sound scarier than it actually is because humanity has been co-working and cohabiting and coexisting with machines already. I don’t think we should hype up the speed but I think it is an inevitable trend coming.
Dan S: There’s some interesting conversations around the idea that the car is an avatar, which is actually pretty accurate. Let’s say you have a blue sedan and your name is Joe and I know you. I see your blue sedan two blocks away, and I go, “Oh, there’s Joe.” I can’t see you. I can’t see your body, but I can see the car. When you go inside of that car, it becomes an avatar for you.
Somehow we like our cars with eyeballs as lights and a big mouth. We like them to be sort of humanoid in a way. When you go into a world where we’re all using a ride-sharing service or using a self-driving car, all this is going to change. That relationship’s going to change. We’re going to lose some of what made our cars our avatars.
Certainly, as we move away from the cookie cutter car into this new world of much more diverse types of vehicles and systems, cars can change. One designer made a car where the whole side of it glowed and had a color to it. In the future it could adapt to your personality and habits externally, too, in terms of color and communication and everything.
Sidney: Think about your dog. You’re having a bad day and he’ll just react to these very, very simple cues. He’ll follow you around or he’ll leave you alone. It’s just sensing a few things and it’s reacting accordingly. That itself can go so far.
I think that in the near term, there’s so much that a machine can accomplish by just responding to very, very subtle cues. This is, by the way, when contextualization comes in. And personalization. Given that now the machine can observe you all the time, it can figure out — essentially, learn — how to respond and what effects it has.
Q; How do you see the next generation of transportation and technology changing people’s daily lives?
Dan S: I talk about a three-tier transportation future. Today most Americans know a two-tier system. Most cities across the metropolitan area use an automobile or a light truck. Then when we want to go further faster, we go to an airport for another tier of transportation, the airplane. We need to create another tier, the higher speed tier, for intercity transportation. We’re going to move to smaller inner-city zones where smaller vehicles as well as bicycles are going to move.
You hear a lot about that with the first and last mile problem, like how people get to the train station from their home. How do I actually get two miles without having to own a car? That can be smaller vehicles. They’re shared transit vehicles or smaller autonomous vehicles. We could even see kind of a Russian doll thing where a small autonomous vehicle could drive right into the house and then it could drive right out and then become a part of a faster tier-two car to go across the region.
It’s just like you would fly on the airplane today. You’d drive to the airport, get on a plane, go across to the other city and then get a car rental.
Sidney: A large part of wanting to change or wanting to improve involves having good, reliable measurement. Look at Fitbits. Having that information, telling you how many steps you walked and things like that, it can be of course be a motivational symbol in addition to giving you good data. I would really love if we had an attentional equivalent of the Fitbit. If somebody is going about their day and there’s a way to sense how inattentive they are or how attentionally focused they are.
Fei Fei Le: There will be and should be a lot of interactions between AI research and security and privacy research. The goal is to make technology that serves people. Not replace, not anything else. I would love to see even more of that. I think it opens up a lot of exciting opportunities. At Stanford we’ve been focusing a lot on AI and benevolent technology. We want to make sure that this movement is making a positive impact to human life
Technology and AI will play significant roles in changing the way we think about driving and transportation. To learn more about the role of AI in cars, check out Toyota’s Concept-i, an exciting glimpse into the future of mobility.